mirror of
https://github.com/vllm-project/vllm.git
synced 2025-10-21 07:13:52 +08:00
Compare commits
3 Commits
memory-lea
...
add-utils
| Author | SHA1 | Date | |
|---|---|---|---|
| 7d092fc32c | |||
| 1a6c27f271 | |||
| 3c6fd286b4 |
@ -5,11 +5,11 @@ import os
|
||||
import sys
|
||||
import zipfile
|
||||
|
||||
# Read the VLLM_MAX_SIZE_MB environment variable, defaulting to 450 MiB
|
||||
# Note that we have 800 MiB quota, please use it wisely.
|
||||
# See https://github.com/pypi/support/issues/6326 .
|
||||
# Read the VLLM_MAX_SIZE_MB environment variable, defaulting to 400 MiB
|
||||
# Note that we have 400 MiB quota, please use it wisely.
|
||||
# See https://github.com/pypi/support/issues/3792 .
|
||||
# Please also sync the value with the one in Dockerfile.
|
||||
VLLM_MAX_SIZE_MB = int(os.environ.get("VLLM_MAX_SIZE_MB", 450))
|
||||
VLLM_MAX_SIZE_MB = int(os.environ.get("VLLM_MAX_SIZE_MB", 400))
|
||||
|
||||
|
||||
def print_top_10_largest_files(zip_file):
|
||||
|
||||
@ -8,8 +8,7 @@ template = """<!DOCTYPE html>
|
||||
<html>
|
||||
<body>
|
||||
<h1>Links for vLLM</h1/>
|
||||
<a href="../{x86_wheel_html_escaped}">{x86_wheel}</a><br/>
|
||||
<a href="../{arm_wheel_html_escaped}">{arm_wheel}</a><br/>
|
||||
<a href="../{wheel_html_escaped}">{wheel}</a><br/>
|
||||
</body>
|
||||
</html>
|
||||
"""
|
||||
@ -22,25 +21,7 @@ filename = os.path.basename(args.wheel)
|
||||
|
||||
with open("index.html", "w") as f:
|
||||
print(f"Generated index.html for {args.wheel}")
|
||||
# sync the abi tag with .buildkite/scripts/upload-wheels.sh
|
||||
if "x86_64" in filename:
|
||||
x86_wheel = filename
|
||||
arm_wheel = filename.replace("x86_64", "aarch64").replace(
|
||||
"manylinux1", "manylinux2014"
|
||||
)
|
||||
elif "aarch64" in filename:
|
||||
x86_wheel = filename.replace("aarch64", "x86_64").replace(
|
||||
"manylinux2014", "manylinux1"
|
||||
)
|
||||
arm_wheel = filename
|
||||
else:
|
||||
raise ValueError(f"Unsupported wheel: {filename}")
|
||||
# cloudfront requires escaping the '+' character
|
||||
f.write(
|
||||
template.format(
|
||||
x86_wheel=x86_wheel,
|
||||
x86_wheel_html_escaped=x86_wheel.replace("+", "%2B"),
|
||||
arm_wheel=arm_wheel,
|
||||
arm_wheel_html_escaped=arm_wheel.replace("+", "%2B"),
|
||||
)
|
||||
template.format(wheel=filename, wheel_html_escaped=filename.replace("+", "%2B"))
|
||||
)
|
||||
|
||||
12
.buildkite/lm-eval-harness/configs/Meta-Llama-3-8B-QQQ.yaml
Normal file
12
.buildkite/lm-eval-harness/configs/Meta-Llama-3-8B-QQQ.yaml
Normal file
@ -0,0 +1,12 @@
|
||||
# For vllm script, with -t option (tensor parallel size).
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-gsm-vllm-baseline.sh -m HandH1998/QQQ-Llama-3-8b-g128 -b 32 -l 1000 -f 5 -t 1
|
||||
model_name: "HandH1998/QQQ-Llama-3-8b-g128"
|
||||
tasks:
|
||||
- name: "gsm8k"
|
||||
metrics:
|
||||
- name: "exact_match,strict-match"
|
||||
value: 0.419
|
||||
- name: "exact_match,flexible-extract"
|
||||
value: 0.416
|
||||
limit: 1000
|
||||
num_fewshot: 5
|
||||
@ -3,3 +3,4 @@ Meta-Llama-3-70B-Instruct.yaml
|
||||
Mixtral-8x7B-Instruct-v0.1.yaml
|
||||
Qwen2-57B-A14-Instruct.yaml
|
||||
DeepSeek-V2-Lite-Chat.yaml
|
||||
Meta-Llama-3-8B-QQQ.yaml
|
||||
|
||||
@ -2,7 +2,7 @@
|
||||
# We can use this script to compute baseline accuracy on GSM for transformers.
|
||||
#
|
||||
# Make sure you have lm-eval-harness installed:
|
||||
# pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d#egg=lm-eval[api]
|
||||
# pip install lm-eval==0.4.4
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
|
||||
@ -3,7 +3,7 @@
|
||||
# We use this for fp8, which HF does not support.
|
||||
#
|
||||
# Make sure you have lm-eval-harness installed:
|
||||
# pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d#egg=lm-eval[api]
|
||||
# pip install lm-eval==0.4.4
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
@ -46,6 +46,6 @@ while getopts "m:b:l:f:t:" OPT; do
|
||||
done
|
||||
|
||||
lm_eval --model vllm \
|
||||
--model_args "pretrained=$MODEL,tensor_parallel_size=$TP_SIZE,add_bos_token=true,trust_remote_code=true,max_model_len=4096" \
|
||||
--model_args "pretrained=$MODEL,tensor_parallel_size=$TP_SIZE,distributed_executor_backend=ray,trust_remote_code=true,max_model_len=4096" \
|
||||
--tasks gsm8k --num_fewshot "$FEWSHOT" --limit "$LIMIT" \
|
||||
--batch_size "$BATCH_SIZE"
|
||||
|
||||
@ -18,14 +18,12 @@ RTOL = 0.08
|
||||
|
||||
def launch_lm_eval(eval_config, tp_size):
|
||||
trust_remote_code = eval_config.get("trust_remote_code", False)
|
||||
max_model_len = eval_config.get("max_model_len", 4096)
|
||||
model_args = (
|
||||
f"pretrained={eval_config['model_name']},"
|
||||
f"tensor_parallel_size={tp_size},"
|
||||
f"enforce_eager=true,"
|
||||
f"add_bos_token=true,"
|
||||
f"trust_remote_code={trust_remote_code},"
|
||||
f"max_model_len={max_model_len}"
|
||||
f"trust_remote_code={trust_remote_code}"
|
||||
)
|
||||
results = lm_eval.simple_evaluate(
|
||||
model="vllm",
|
||||
|
||||
@ -7,7 +7,7 @@ This directory contains two sets of benchmark for vllm.
|
||||
- Performance benchmark: benchmark vllm's performance under various workload, for **developers** to gain clarity on whether their PR improves/degrades vllm's performance
|
||||
- Nightly benchmark: compare vllm's performance against alternatives (tgi, trt-llm and lmdeploy), for **the public** to know when to choose vllm.
|
||||
|
||||
See [vLLM performance dashboard](https://hud.pytorch.org/benchmark/llms?repoName=vllm-project%2Fvllm) for the latest performance benchmark results and [vLLM GitHub README](https://github.com/vllm-project/vllm/blob/main/README.md) for latest nightly benchmark results.
|
||||
See [vLLM performance dashboard](https://perf.vllm.ai) for the latest performance benchmark results and [vLLM GitHub README](https://github.com/vllm-project/vllm/blob/main/README.md) for latest nightly benchmark results.
|
||||
|
||||
## Performance benchmark quick overview
|
||||
|
||||
@ -28,7 +28,6 @@ See [vLLM performance dashboard](https://hud.pytorch.org/benchmark/llms?repoName
|
||||
## Trigger the benchmark
|
||||
|
||||
Performance benchmark will be triggered when:
|
||||
|
||||
- A PR being merged into vllm.
|
||||
- Every commit for those PRs with `perf-benchmarks` label AND `ready` label.
|
||||
|
||||
@ -39,7 +38,6 @@ bash .buildkite/nightly-benchmarks/scripts/run-performance-benchmarks.sh
|
||||
```
|
||||
|
||||
Runtime environment variables:
|
||||
|
||||
- `ON_CPU`: set the value to '1' on Intel® Xeon® Processors. Default value is 0.
|
||||
- `SERVING_JSON`: JSON file to use for the serving tests. Default value is empty string (use default file).
|
||||
- `LATENCY_JSON`: JSON file to use for the latency tests. Default value is empty string (use default file).
|
||||
@ -48,14 +46,12 @@ Runtime environment variables:
|
||||
- `REMOTE_PORT`: Port for the remote vLLM service to benchmark. Default value is empty string.
|
||||
|
||||
Nightly benchmark will be triggered when:
|
||||
|
||||
- Every commit for those PRs with `perf-benchmarks` label and `nightly-benchmarks` label.
|
||||
|
||||
## Performance benchmark details
|
||||
|
||||
See [performance-benchmarks-descriptions.md](performance-benchmarks-descriptions.md) for detailed descriptions, and use `tests/latency-tests.json`, `tests/throughput-tests.json`, `tests/serving-tests.json` to configure the test cases.
|
||||
> NOTE: For Intel® Xeon® Processors, use `tests/latency-tests-cpu.json`, `tests/throughput-tests-cpu.json`, `tests/serving-tests-cpu.json` instead.
|
||||
>
|
||||
### Latency test
|
||||
|
||||
Here is an example of one test inside `latency-tests.json`:
|
||||
@ -78,7 +74,7 @@ Here is an example of one test inside `latency-tests.json`:
|
||||
In this example:
|
||||
|
||||
- The `test_name` attributes is a unique identifier for the test. In `latency-tests.json`, it must start with `latency_`.
|
||||
- The `parameters` attribute control the command line arguments to be used for `vllm bench latency`. Note that please use underline `_` instead of the dash `-` when specifying the command line arguments, and `run-performance-benchmarks.sh` will convert the underline to dash when feeding the arguments to `vllm bench latency`. For example, the corresponding command line arguments for `vllm bench latency` will be `--model meta-llama/Meta-Llama-3-8B --tensor-parallel-size 1 --load-format dummy --num-iters-warmup 5 --num-iters 15`
|
||||
- The `parameters` attribute control the command line arguments to be used for `benchmark_latency.py`. Note that please use underline `_` instead of the dash `-` when specifying the command line arguments, and `run-performance-benchmarks.sh` will convert the underline to dash when feeding the arguments to `benchmark_latency.py`. For example, the corresponding command line arguments for `benchmark_latency.py` will be `--model meta-llama/Meta-Llama-3-8B --tensor-parallel-size 1 --load-format dummy --num-iters-warmup 5 --num-iters 15`
|
||||
|
||||
Note that the performance numbers are highly sensitive to the value of the parameters. Please make sure the parameters are set correctly.
|
||||
|
||||
@ -86,13 +82,13 @@ WARNING: The benchmarking script will save json results by itself, so please do
|
||||
|
||||
### Throughput test
|
||||
|
||||
The tests are specified in `throughput-tests.json`. The syntax is similar to `latency-tests.json`, except for that the parameters will be fed forward to `vllm bench throughput`.
|
||||
The tests are specified in `throughput-tests.json`. The syntax is similar to `latency-tests.json`, except for that the parameters will be fed forward to `benchmark_throughput.py`.
|
||||
|
||||
The number of this test is also stable -- a slight change on the value of this number might vary the performance numbers by a lot.
|
||||
|
||||
### Serving test
|
||||
|
||||
We test the throughput by using `vllm bench serve` with request rate = inf to cover the online serving overhead. The corresponding parameters are in `serving-tests.json`, and here is an example:
|
||||
We test the throughput by using `benchmark_serving.py` with request rate = inf to cover the online serving overhead. The corresponding parameters are in `serving-tests.json`, and here is an example:
|
||||
|
||||
```json
|
||||
[
|
||||
@ -104,6 +100,7 @@ We test the throughput by using `vllm bench serve` with request rate = inf to co
|
||||
"tensor_parallel_size": 1,
|
||||
"swap_space": 16,
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
@ -121,8 +118,8 @@ Inside this example:
|
||||
|
||||
- The `test_name` attribute is also a unique identifier for the test. It must start with `serving_`.
|
||||
- The `server-parameters` includes the command line arguments for vLLM server.
|
||||
- The `client-parameters` includes the command line arguments for `vllm bench serve`.
|
||||
- The `qps_list` controls the list of qps for test. It will be used to configure the `--request-rate` parameter in `vllm bench serve`
|
||||
- The `client-parameters` includes the command line arguments for `benchmark_serving.py`.
|
||||
- The `qps_list` controls the list of qps for test. It will be used to configure the `--request-rate` parameter in `benchmark_serving.py`
|
||||
|
||||
The number of this test is less stable compared to the delay and latency benchmarks (due to randomized sharegpt dataset sampling inside `benchmark_serving.py`), but a large change on this number (e.g. 5% change) still vary the output greatly.
|
||||
|
||||
@ -138,20 +135,27 @@ The raw benchmarking results (in the format of json files) are in the `Artifacts
|
||||
|
||||
The `compare-json-results.py` helps to compare benchmark results JSON files converted using `convert-results-json-to-markdown.py`.
|
||||
When run, benchmark script generates results under `benchmark/results` folder, along with the `benchmark_results.md` and `benchmark_results.json`.
|
||||
`compare-json-results.py` compares two `benchmark_results.json` files and provides performance ratio e.g. for Output Tput, Median TTFT and Median TPOT.
|
||||
If only one benchmark_results.json is passed, `compare-json-results.py` compares different TP and PP configurations in the benchmark_results.json instead.
|
||||
`compare-json-results.py` compares two `benchmark_results.json` files and provides performance ratio e.g. for Output Tput, Median TTFT and Median TPOT.
|
||||
|
||||
Here is an example using the script to compare result_a and result_b with Model, Dataset name, input/output length, max concurrency and qps.
|
||||
Here is an example using the script to compare result_a and result_b without detail test name.
|
||||
`python3 compare-json-results.py -f results_a/benchmark_results.json -f results_b/benchmark_results.json --ignore_test_name`
|
||||
|
||||
| | results_a/benchmark_results.json | results_b/benchmark_results.json | perf_ratio |
|
||||
|----|----------------------------------------|----------------------------------------|----------|
|
||||
| 0 | 142.633982 | 156.526018 | 1.097396 |
|
||||
| 1 | 241.620334 | 294.018783 | 1.216863 |
|
||||
| 2 | 218.298905 | 262.664916 | 1.203235 |
|
||||
| 3 | 242.743860 | 299.816190 | 1.235113 |
|
||||
|
||||
Here is an example using the script to compare result_a and result_b with detail test name.
|
||||
`python3 compare-json-results.py -f results_a/benchmark_results.json -f results_b/benchmark_results.json`
|
||||
|
||||
| | Model | Dataset Name | Input Len | Output Len | # of max concurrency | qps | results_a/benchmark_results.json | results_b/benchmark_results.json | perf_ratio |
|
||||
|----|---------------------------------------|--------|-----|-----|------|-----|-----------|----------|----------|
|
||||
| 0 | meta-llama/Meta-Llama-3.1-8B-Instruct | random | 128 | 128 | 1000 | 1 | 142.633982 | 156.526018 | 1.097396 |
|
||||
| 1 | meta-llama/Meta-Llama-3.1-8B-Instruct | random | 128 | 128 | 1000 | inf| 241.620334 | 294.018783 | 1.216863 |
|
||||
|
||||
A comparison diagram will be generated below the table.
|
||||
Here is an example to compare between 96c/results_gnr_96c_091_tp2pp3 and 128c/results_gnr_128c_091_tp2pp3
|
||||
<img width="1886" height="828" alt="image" src="https://github.com/user-attachments/assets/c02a43ef-25d0-4fd6-90e5-2169a28682dd" />
|
||||
| | results_a/benchmark_results.json_name | results_a/benchmark_results.json | results_b/benchmark_results.json_name | results_b/benchmark_results.json | perf_ratio |
|
||||
|---|---------------------------------------------|----------------------------------------|---------------------------------------------|----------------------------------------|----------|
|
||||
| 0 | serving_llama8B_tp1_sharegpt_qps_1 | 142.633982 | serving_llama8B_tp1_sharegpt_qps_1 | 156.526018 | 1.097396 |
|
||||
| 1 | serving_llama8B_tp1_sharegpt_qps_16 | 241.620334 | serving_llama8B_tp1_sharegpt_qps_16 | 294.018783 | 1.216863 |
|
||||
| 2 | serving_llama8B_tp1_sharegpt_qps_4 | 218.298905 | serving_llama8B_tp1_sharegpt_qps_4 | 262.664916 | 1.203235 |
|
||||
| 3 | serving_llama8B_tp1_sharegpt_qps_inf | 242.743860 | serving_llama8B_tp1_sharegpt_qps_inf | 299.816190 | 1.235113 |
|
||||
| 4 | serving_llama8B_tp2_random_1024_128_qps_1 | 96.613390 | serving_llama8B_tp4_random_1024_128_qps_1 | 108.404853 | 1.122048 |
|
||||
|
||||
## Nightly test details
|
||||
|
||||
@ -160,9 +164,9 @@ See [nightly-descriptions.md](nightly-descriptions.md) for the detailed descript
|
||||
### Workflow
|
||||
|
||||
- The [nightly-pipeline.yaml](nightly-pipeline.yaml) specifies the docker containers for different LLM serving engines.
|
||||
- Inside each container, we run [scripts/run-nightly-benchmarks.sh](scripts/run-nightly-benchmarks.sh), which will probe the serving engine of the current container.
|
||||
- The `scripts/run-nightly-benchmarks.sh` will parse the workload described in [nightly-tests.json](tests/nightly-tests.json) and launch the right benchmark for the specified serving engine via `scripts/launch-server.sh`.
|
||||
- At last, we run [scripts/summary-nightly-results.py](scripts/summary-nightly-results.py) to collect and plot the final benchmarking results, and update the results to buildkite.
|
||||
- Inside each container, we run [run-nightly-suite.sh](run-nightly-suite.sh), which will probe the serving engine of the current container.
|
||||
- The `run-nightly-suite.sh` will redirect the request to `tests/run-[llm serving engine name]-nightly.sh`, which parses the workload described in [nightly-tests.json](tests/nightly-tests.json) and performs the benchmark.
|
||||
- At last, we run [scripts/plot-nightly-results.py](scripts/plot-nightly-results.py) to collect and plot the final benchmarking results, and update the results to buildkite.
|
||||
|
||||
### Nightly tests
|
||||
|
||||
@ -172,6 +176,6 @@ In [nightly-tests.json](tests/nightly-tests.json), we include the command line a
|
||||
|
||||
The docker containers for benchmarking are specified in `nightly-pipeline.yaml`.
|
||||
|
||||
WARNING: the docker versions are HARD-CODED and SHOULD BE ALIGNED WITH `nightly-descriptions.md`. The docker versions need to be hard-coded as there are several version-specific bug fixes inside `scripts/run-nightly-benchmarks.sh` and `scripts/launch-server.sh`.
|
||||
WARNING: the docker versions are HARD-CODED and SHOULD BE ALIGNED WITH `nightly-descriptions.md`. The docker versions need to be hard-coded as there are several version-specific bug fixes inside `tests/run-[llm serving engine name]-nightly.sh`.
|
||||
|
||||
WARNING: populating `trt-llm` to latest version is not easy, as it requires updating several protobuf files in [tensorrt-demo](https://github.com/neuralmagic/tensorrt-demo.git).
|
||||
|
||||
@ -1,4 +1,3 @@
|
||||
# Nightly benchmark annotation
|
||||
|
||||
## Description
|
||||
|
||||
@ -14,15 +13,15 @@ Please download the visualization scripts in the post
|
||||
|
||||
- Find the docker we use in `benchmarking pipeline`
|
||||
- Deploy the docker, and inside the docker:
|
||||
- Download `nightly-benchmarks.zip`.
|
||||
- In the same folder, run the following code:
|
||||
- Download `nightly-benchmarks.zip`.
|
||||
- In the same folder, run the following code:
|
||||
|
||||
```bash
|
||||
export HF_TOKEN=<your HF token>
|
||||
apt update
|
||||
apt install -y git
|
||||
unzip nightly-benchmarks.zip
|
||||
VLLM_SOURCE_CODE_LOC=./ bash .buildkite/nightly-benchmarks/scripts/run-nightly-benchmarks.sh
|
||||
```
|
||||
```bash
|
||||
export HF_TOKEN=<your HF token>
|
||||
apt update
|
||||
apt install -y git
|
||||
unzip nightly-benchmarks.zip
|
||||
VLLM_SOURCE_CODE_LOC=./ bash .buildkite/nightly-benchmarks/scripts/run-nightly-benchmarks.sh
|
||||
```
|
||||
|
||||
And the results will be inside `./benchmarks/results`.
|
||||
|
||||
@ -13,25 +13,25 @@ Latest reproduction guilde: [github issue link](https://github.com/vllm-project/
|
||||
## Setup
|
||||
|
||||
- Docker images:
|
||||
- vLLM: `vllm/vllm-openai:v0.6.2`
|
||||
- SGLang: `lmsysorg/sglang:v0.3.2-cu121`
|
||||
- LMDeploy: `openmmlab/lmdeploy:v0.6.1-cu12`
|
||||
- TensorRT-LLM: `nvcr.io/nvidia/tritonserver:24.07-trtllm-python-py3`
|
||||
- *NOTE: we use r24.07 as the current implementation only works for this version. We are going to bump this up.*
|
||||
- Check [nightly-pipeline.yaml](nightly-pipeline.yaml) for the concrete docker images, specs and commands we use for the benchmark.
|
||||
- vLLM: `vllm/vllm-openai:v0.6.2`
|
||||
- SGLang: `lmsysorg/sglang:v0.3.2-cu121`
|
||||
- LMDeploy: `openmmlab/lmdeploy:v0.6.1-cu12`
|
||||
- TensorRT-LLM: `nvcr.io/nvidia/tritonserver:24.07-trtllm-python-py3`
|
||||
- *NOTE: we uses r24.07 as the current implementation only works for this version. We are going to bump this up.*
|
||||
- Check [nightly-pipeline.yaml](nightly-pipeline.yaml) for the concrete docker images, specs and commands we use for the benchmark.
|
||||
- Hardware
|
||||
- 8x Nvidia A100 GPUs
|
||||
- 8x Nvidia A100 GPUs
|
||||
- Workload:
|
||||
- Dataset
|
||||
- ShareGPT dataset
|
||||
- Prefill-heavy dataset (in average 462 input tokens, 16 tokens as output)
|
||||
- Decode-heavy dataset (in average 462 input tokens, 256 output tokens)
|
||||
- Check [nightly-tests.json](tests/nightly-tests.json) for the concrete configuration of datasets we use.
|
||||
- Models: llama-3 8B, llama-3 70B.
|
||||
- We do not use llama 3.1 as it is incompatible with trt-llm r24.07. ([issue](https://github.com/NVIDIA/TensorRT-LLM/issues/2105)).
|
||||
- Average QPS (query per second): 2, 4, 8, 16, 32 and inf.
|
||||
- Queries are randomly sampled, and arrival patterns are determined via Poisson process, but all with fixed random seed.
|
||||
- Evaluation metrics: Throughput (higher the better), TTFT (time to the first token, lower the better), ITL (inter-token latency, lower the better).
|
||||
- Dataset
|
||||
- ShareGPT dataset
|
||||
- Prefill-heavy dataset (in average 462 input tokens, 16 tokens as output)
|
||||
- Decode-heavy dataset (in average 462 input tokens, 256 output tokens)
|
||||
- Check [nightly-tests.json](tests/nightly-tests.json) for the concrete configuration of datasets we use.
|
||||
- Models: llama-3 8B, llama-3 70B.
|
||||
- We do not use llama 3.1 as it is incompatible with trt-llm r24.07. ([issue](https://github.com/NVIDIA/TensorRT-LLM/issues/2105)).
|
||||
- Average QPS (query per second): 2, 4, 8, 16, 32 and inf.
|
||||
- Queries are randomly sampled, and arrival patterns are determined via Poisson process, but all with fixed random seed.
|
||||
- Evaluation metrics: Throughput (higher the better), TTFT (time to the first token, lower the better), ITL (inter-token latency, lower the better).
|
||||
|
||||
## Known issues
|
||||
|
||||
|
||||
@ -1,4 +1,3 @@
|
||||
# Performance benchmarks descriptions
|
||||
|
||||
## Latency tests
|
||||
|
||||
|
||||
@ -1,202 +1,33 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import argparse
|
||||
import json
|
||||
import os
|
||||
from importlib import util
|
||||
|
||||
import pandas as pd
|
||||
|
||||
plotly_found = util.find_spec("plotly.express") is not None
|
||||
|
||||
|
||||
def compare_data_columns(
|
||||
files, name_column, data_column, info_cols, drop_column, debug=False
|
||||
files, name_column, data_column, drop_column, ignore_test_name=False
|
||||
):
|
||||
"""
|
||||
Align concatenation by keys derived from info_cols instead of row order.
|
||||
- Pick one canonical key list: subset of info_cols present in ALL files.
|
||||
- For each file: set index to those keys, aggregate duplicates
|
||||
- (mean for metric, first for names).
|
||||
- Concat along axis=1 (indexes align), then reset_index so callers can
|
||||
- group by columns.
|
||||
- If --debug, add a <file_label>_name column per file.
|
||||
"""
|
||||
print("\ncompare_data_column:", data_column)
|
||||
|
||||
print("\ncompare_data_column: " + data_column)
|
||||
frames = []
|
||||
raw_data_cols = []
|
||||
compare_frames = []
|
||||
|
||||
# 1) choose a canonical key list from info_cols that exists in ALL files
|
||||
cols_per_file = []
|
||||
for f in files:
|
||||
try:
|
||||
df_tmp = pd.read_json(f, orient="records")
|
||||
except Exception as err:
|
||||
raise ValueError(f"Failed to read {f}") from err
|
||||
cols_per_file.append(set(df_tmp.columns))
|
||||
|
||||
key_cols = [c for c in info_cols if all(c in cset for cset in cols_per_file)]
|
||||
if not key_cols:
|
||||
# soft fallback: use any info_cols present in the first file
|
||||
key_cols = [c for c in info_cols if c in list(cols_per_file[0])]
|
||||
if not key_cols:
|
||||
raise ValueError(
|
||||
"No common key columns found from info_cols across the input files."
|
||||
)
|
||||
|
||||
# 2) build a single "meta" block (keys as columns) once, aligned by the key index
|
||||
meta_added = False
|
||||
|
||||
for file in files:
|
||||
df = pd.read_json(file, orient="records")
|
||||
|
||||
# Keep rows that actually have the compared metric (same as original behavior)
|
||||
if drop_column in df.columns:
|
||||
df = df.dropna(subset=[drop_column], ignore_index=True)
|
||||
|
||||
# Stabilize numeric key columns (harmless if missing)
|
||||
for c in (
|
||||
"Input Len",
|
||||
"Output Len",
|
||||
"TP Size",
|
||||
"PP Size",
|
||||
"# of max concurrency.",
|
||||
"qps",
|
||||
):
|
||||
if c in df.columns:
|
||||
df[c] = pd.to_numeric(df[c], errors="coerce")
|
||||
|
||||
# Ensure all key columns exist
|
||||
for c in key_cols:
|
||||
if c not in df.columns:
|
||||
df[c] = pd.NA
|
||||
|
||||
# Set index = key_cols and aggregate duplicates → unique MultiIndex
|
||||
df_idx = df.set_index(key_cols, drop=False)
|
||||
|
||||
# meta (key columns), unique per key
|
||||
meta = df_idx[key_cols]
|
||||
if not meta.index.is_unique:
|
||||
meta = meta.groupby(level=key_cols, dropna=False).first()
|
||||
|
||||
# metric series for this file, aggregated to one row per key
|
||||
file_label = "/".join(file.split("/")[:-1]) or os.path.basename(file)
|
||||
s = df_idx[data_column]
|
||||
if not s.index.is_unique:
|
||||
s = s.groupby(level=key_cols, dropna=False).mean()
|
||||
s.name = file_label # column label like original
|
||||
|
||||
# add meta once (from first file) so keys are the leftmost columns
|
||||
if not meta_added:
|
||||
frames.append(meta)
|
||||
meta_added = True
|
||||
|
||||
# (NEW) debug: aligned test-name column per file
|
||||
if debug and name_column in df_idx.columns:
|
||||
name_s = df_idx[name_column]
|
||||
if not name_s.index.is_unique:
|
||||
name_s = name_s.groupby(level=key_cols, dropna=False).first()
|
||||
name_s.name = f"{file_label}_name"
|
||||
frames.append(name_s)
|
||||
|
||||
frames.append(s)
|
||||
raw_data_cols.append(file_label)
|
||||
compare_frames.append(s)
|
||||
|
||||
# Generalize ratio: for any file N>=2, add ratio (fileN / file1)
|
||||
data_df = pd.read_json(file)
|
||||
serving_df = data_df.dropna(subset=[drop_column], ignore_index=True)
|
||||
if ignore_test_name is False:
|
||||
serving_df = serving_df.rename(columns={name_column: file + "_name"})
|
||||
frames.append(serving_df[file + "_name"])
|
||||
serving_df = serving_df.rename(columns={data_column: file})
|
||||
frames.append(serving_df[file])
|
||||
compare_frames.append(serving_df[file])
|
||||
if len(compare_frames) >= 2:
|
||||
base = compare_frames[0]
|
||||
current = compare_frames[-1]
|
||||
ratio = current / base
|
||||
ratio = ratio.mask(base == 0) # avoid inf when baseline is 0
|
||||
ratio.name = f"Ratio 1 vs {len(compare_frames)}"
|
||||
frames.append(ratio)
|
||||
# Compare numbers among two files
|
||||
ratio_df = compare_frames[1] / compare_frames[0]
|
||||
frames.append(ratio_df)
|
||||
compare_frames.pop(1)
|
||||
|
||||
# 4) concat on columns with aligned MultiIndex;
|
||||
# then reset_index to return keys as columns
|
||||
concat_df = pd.concat(frames, axis=1)
|
||||
concat_df = concat_df.reset_index(drop=True).reset_index()
|
||||
if "index" in concat_df.columns:
|
||||
concat_df = concat_df.drop(columns=["index"])
|
||||
|
||||
# Ensure key/info columns appear first (in your info_cols order)
|
||||
front = [c for c in info_cols if c in concat_df.columns]
|
||||
rest = [c for c in concat_df.columns if c not in front]
|
||||
concat_df = concat_df[front + rest]
|
||||
|
||||
print(raw_data_cols)
|
||||
return concat_df, raw_data_cols
|
||||
|
||||
|
||||
def split_json_by_tp_pp(
|
||||
input_file: str = "benchmark_results.json", output_root: str = "."
|
||||
) -> list[str]:
|
||||
"""
|
||||
Split a benchmark JSON into separate folders by (TP Size, PP Size).
|
||||
|
||||
Creates: <output_root>/tp{TP}_pp{PP}/benchmark_results.json
|
||||
Returns: list of file paths written.
|
||||
"""
|
||||
# Load JSON data into DataFrame
|
||||
with open(input_file, encoding="utf-8") as f:
|
||||
data = json.load(f)
|
||||
|
||||
# If the JSON is a dict with a list under common keys, use that list
|
||||
if isinstance(data, dict):
|
||||
for key in ("results", "serving_results", "benchmarks", "data"):
|
||||
if isinstance(data.get(key), list):
|
||||
data = data[key]
|
||||
break
|
||||
|
||||
df = pd.DataFrame(data)
|
||||
|
||||
# Keep only "serving" tests
|
||||
name_col = next(
|
||||
(c for c in ["Test name", "test_name", "Test Name"] if c in df.columns), None
|
||||
)
|
||||
if name_col:
|
||||
df = df[
|
||||
df[name_col].astype(str).str.contains(r"serving", case=False, na=False)
|
||||
].copy()
|
||||
|
||||
# Handle alias column names
|
||||
rename_map = {
|
||||
"tp_size": "TP Size",
|
||||
"tensor_parallel_size": "TP Size",
|
||||
"pp_size": "PP Size",
|
||||
"pipeline_parallel_size": "PP Size",
|
||||
}
|
||||
df.rename(
|
||||
columns={k: v for k, v in rename_map.items() if k in df.columns}, inplace=True
|
||||
)
|
||||
|
||||
# Ensure TP/PP columns exist (default to 1 if missing)
|
||||
if "TP Size" not in df.columns:
|
||||
df["TP Size"] = 1
|
||||
if "PP Size" not in df.columns:
|
||||
df["PP Size"] = 1
|
||||
|
||||
# make sure TP/PP are numeric ints with no NaN
|
||||
df["TP Size"] = (
|
||||
pd.to_numeric(df.get("TP Size", 1), errors="coerce").fillna(1).astype(int)
|
||||
)
|
||||
df["PP Size"] = (
|
||||
pd.to_numeric(df.get("PP Size", 1), errors="coerce").fillna(1).astype(int)
|
||||
)
|
||||
|
||||
# Split into separate folders
|
||||
saved_paths: list[str] = []
|
||||
for (tp, pp), group_df in df.groupby(["TP Size", "PP Size"], dropna=False):
|
||||
folder_name = os.path.join(output_root, f"tp{int(tp)}_pp{int(pp)}")
|
||||
os.makedirs(folder_name, exist_ok=True)
|
||||
filepath = os.path.join(folder_name, "benchmark_results.json")
|
||||
group_df.to_json(filepath, orient="records", indent=2, force_ascii=False)
|
||||
print(f"Saved: {filepath}")
|
||||
saved_paths.append(filepath)
|
||||
|
||||
return saved_paths
|
||||
return concat_df
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
@ -205,103 +36,31 @@ if __name__ == "__main__":
|
||||
"-f", "--file", action="append", type=str, help="input file name"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--debug", action="store_true", help="show all information for debugging"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--plot",
|
||||
action=argparse.BooleanOptionalAction,
|
||||
default=True,
|
||||
help="plot perf diagrams or not --no-plot --plot",
|
||||
)
|
||||
parser.add_argument(
|
||||
"-x",
|
||||
"--xaxis",
|
||||
type=str,
|
||||
default="# of max concurrency.",
|
||||
help="column name to use as X Axis in comparison graph",
|
||||
"--ignore_test_name", action="store_true", help="ignore_test_name or not"
|
||||
)
|
||||
args = parser.parse_args()
|
||||
files = args.file
|
||||
print("comparing : " + ", ".join(files))
|
||||
|
||||
drop_column = "P99"
|
||||
name_column = "Test name"
|
||||
info_cols = [
|
||||
"Model",
|
||||
"Dataset Name",
|
||||
"Input Len",
|
||||
"Output Len",
|
||||
"TP Size",
|
||||
"PP Size",
|
||||
"# of max concurrency.",
|
||||
"qps",
|
||||
]
|
||||
data_cols_to_compare = ["Output Tput (tok/s)", "Median TTFT (ms)", "Median"]
|
||||
html_msgs_for_data_cols = [
|
||||
"Compare Output Tokens /n",
|
||||
"Median TTFT /n",
|
||||
"Median TPOT /n",
|
||||
]
|
||||
|
||||
if len(args.file) == 1:
|
||||
files = split_json_by_tp_pp(args.file[0], output_root="splits")
|
||||
info_cols = [c for c in info_cols if c not in ("TP Size", "PP Size")]
|
||||
else:
|
||||
files = args.file
|
||||
print("comparing : " + ", ".join(files))
|
||||
debug = args.debug
|
||||
plot = args.plot
|
||||
# For Plot feature, assign y axis from one of info_cols
|
||||
y_axis_index = info_cols.index(args.xaxis) if args.xaxis in info_cols else 6
|
||||
ignore_test_name = args.ignore_test_name
|
||||
with open("perf_comparison.html", "w") as text_file:
|
||||
for i in range(len(data_cols_to_compare)):
|
||||
output_df, raw_data_cols = compare_data_columns(
|
||||
output_df = compare_data_columns(
|
||||
files,
|
||||
name_column,
|
||||
data_cols_to_compare[i],
|
||||
info_cols,
|
||||
drop_column,
|
||||
debug=debug,
|
||||
ignore_test_name=ignore_test_name,
|
||||
)
|
||||
|
||||
# For Plot feature, insert y axis from one of info_cols
|
||||
raw_data_cols.insert(0, info_cols[y_axis_index])
|
||||
|
||||
filtered_info_cols = info_cols[:-2]
|
||||
existing_group_cols = [
|
||||
c for c in filtered_info_cols if c in output_df.columns
|
||||
]
|
||||
if not existing_group_cols:
|
||||
raise ValueError(
|
||||
f"No valid group-by columns "
|
||||
f"Expected subset: {filtered_info_cols}, "
|
||||
f"but DataFrame has: {list(output_df.columns)}"
|
||||
)
|
||||
output_df_sorted = output_df.sort_values(by=existing_group_cols)
|
||||
output_groups = output_df_sorted.groupby(existing_group_cols, dropna=False)
|
||||
for name, group in output_groups:
|
||||
html = group.to_html()
|
||||
text_file.write(html_msgs_for_data_cols[i])
|
||||
text_file.write(html)
|
||||
|
||||
if plot and plotly_found:
|
||||
import plotly.express as px
|
||||
|
||||
df = group[raw_data_cols]
|
||||
df_sorted = df.sort_values(by=info_cols[y_axis_index])
|
||||
# Melt DataFrame for plotting
|
||||
df_melted = df_sorted.melt(
|
||||
id_vars=info_cols[y_axis_index],
|
||||
var_name="Configuration",
|
||||
value_name=data_cols_to_compare[i],
|
||||
)
|
||||
title = data_cols_to_compare[i] + " vs " + info_cols[y_axis_index]
|
||||
# Create Plotly line chart
|
||||
fig = px.line(
|
||||
df_melted,
|
||||
x=info_cols[y_axis_index],
|
||||
y=data_cols_to_compare[i],
|
||||
color="Configuration",
|
||||
title=title,
|
||||
markers=True,
|
||||
)
|
||||
# Export to HTML
|
||||
text_file.write(fig.to_html(full_html=True, include_plotlyjs="cdn"))
|
||||
print(output_df)
|
||||
html = output_df.to_html()
|
||||
text_file.write(html_msgs_for_data_cols[i])
|
||||
text_file.write(html)
|
||||
|
||||
@ -1,19 +1,17 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import os
|
||||
import shlex
|
||||
from importlib import util
|
||||
from pathlib import Path
|
||||
from typing import Any
|
||||
|
||||
import pandas as pd
|
||||
import psutil
|
||||
import regex as re
|
||||
from tabulate import tabulate
|
||||
|
||||
results_folder = Path("results/")
|
||||
|
||||
# latency results and the keys that will be printed into markdown
|
||||
latency_results = []
|
||||
latency_column_mapping = {
|
||||
@ -44,22 +42,13 @@ throughput_results_column_mapping = {
|
||||
serving_results = []
|
||||
serving_column_mapping = {
|
||||
"test_name": "Test name",
|
||||
"model_id": "Model",
|
||||
"dataset_name": "Dataset Name",
|
||||
"input_len": "Input Len",
|
||||
"output_len": "Output Len",
|
||||
"tp_size": "TP Size",
|
||||
"pp_size": "PP Size",
|
||||
"dtype": "dtype",
|
||||
"gpu_type": "GPU",
|
||||
"completed": "# of req.",
|
||||
"qps": "qps",
|
||||
"max_concurrency": "# of max concurrency.",
|
||||
"request_throughput": "Tput (req/s)",
|
||||
"total_token_throughput": "Total Token Tput (tok/s)",
|
||||
"output_throughput": "Output Tput (tok/s)",
|
||||
# "total_input_tokens": "Total input tokens",
|
||||
# "total_output_tokens": "Total output tokens",
|
||||
"total_input_tokens": "Total input tokens",
|
||||
"total_output_tokens": "Total output tokens",
|
||||
"mean_ttft_ms": "Mean TTFT (ms)",
|
||||
"median_ttft_ms": "Median TTFT (ms)",
|
||||
"p99_ttft_ms": "P99 TTFT (ms)",
|
||||
@ -104,111 +93,15 @@ def get_size_with_unit(bytes, suffix="B"):
|
||||
bytes /= factor
|
||||
|
||||
|
||||
def _coerce(val: str) -> Any:
|
||||
"""Best-effort type coercion from string to Python types."""
|
||||
low = val.lower()
|
||||
if low == "null":
|
||||
return None
|
||||
if low == "true":
|
||||
return True
|
||||
if low == "false":
|
||||
return False
|
||||
# integers
|
||||
if re.fullmatch(r"[+-]?\d+", val):
|
||||
try:
|
||||
return int(val)
|
||||
except ValueError:
|
||||
pass
|
||||
# floats (keep 'inf'/'-inf'/'nan' as strings)
|
||||
if re.fullmatch(r"[+-]?\d*\.\d+", val):
|
||||
try:
|
||||
return float(val)
|
||||
except ValueError:
|
||||
pass
|
||||
return val
|
||||
|
||||
|
||||
def parse_client_command(cmd: str) -> dict[str, Any]:
|
||||
"""Parse the client_command shell string into {executable, script, args}."""
|
||||
toks = shlex.split(cmd)
|
||||
if len(toks) < 2:
|
||||
raise ValueError("client_command must include an executable and a script")
|
||||
executable, script = toks[0], toks[1]
|
||||
args: dict[str, Any] = {}
|
||||
|
||||
i = 2
|
||||
while i < len(toks):
|
||||
t = toks[i]
|
||||
if t.startswith("--"):
|
||||
# --key=value or --key (value) or boolean flag
|
||||
if "=" in t:
|
||||
key, val = t.split("=", 1)
|
||||
if key == "--metadata":
|
||||
md = {}
|
||||
if val:
|
||||
if "=" in val:
|
||||
k, v = val.split("=", 1)
|
||||
md[k] = _coerce(v)
|
||||
else:
|
||||
md[val] = True
|
||||
args[key] = md
|
||||
else:
|
||||
args[key] = _coerce(val)
|
||||
i += 1
|
||||
continue
|
||||
|
||||
key = t
|
||||
|
||||
# Special: consume metadata k=v pairs until next --flag
|
||||
if key == "--metadata":
|
||||
i += 1
|
||||
md = {}
|
||||
while i < len(toks) and not toks[i].startswith("--"):
|
||||
pair = toks[i]
|
||||
if "=" in pair:
|
||||
k, v = pair.split("=", 1)
|
||||
md[k] = _coerce(v)
|
||||
else:
|
||||
md[pair] = True
|
||||
i += 1
|
||||
args[key] = md
|
||||
continue
|
||||
|
||||
# Standard: check if next token is a value (not a flag)
|
||||
if i + 1 < len(toks) and not toks[i + 1].startswith("--"):
|
||||
args[key] = _coerce(toks[i + 1])
|
||||
i += 2
|
||||
else:
|
||||
# lone flag -> True
|
||||
args[key] = True
|
||||
i += 1
|
||||
else:
|
||||
# unexpected positional; skip
|
||||
i += 1
|
||||
|
||||
return {"executable": executable, "script": script, "args": args}
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument(
|
||||
"-r",
|
||||
"--result",
|
||||
type=str,
|
||||
default="results",
|
||||
help="Folder name for benchmark output results.",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
results_folder = Path(args.result)
|
||||
if not results_folder.exists():
|
||||
raise FileNotFoundError(f"results folder does not exist: {results_folder}")
|
||||
# collect results
|
||||
for test_file in results_folder.glob("*.json"):
|
||||
with open(test_file) as f:
|
||||
raw_result = json.loads(f.read())
|
||||
|
||||
if "serving" in str(test_file):
|
||||
# this result is generated via `vllm bench serve` command
|
||||
# this result is generated via `benchmark_serving.py`
|
||||
|
||||
# attach the benchmarking command to raw_result
|
||||
try:
|
||||
with open(test_file.with_suffix(".commands")) as f:
|
||||
@ -216,50 +109,18 @@ if __name__ == "__main__":
|
||||
except OSError as e:
|
||||
print(e)
|
||||
continue
|
||||
# Parse Server Command Arg
|
||||
out: dict[str, Any] = {
|
||||
"server_command": parse_client_command(command["server_command"])
|
||||
}
|
||||
parse_args = [
|
||||
"--tensor-parallel-size",
|
||||
"--pipeline-parallel-size",
|
||||
"--dtype",
|
||||
]
|
||||
col_mapping = ["tp_size", "pp_size", "dtype"]
|
||||
for index, arg in enumerate(parse_args):
|
||||
if arg in out["server_command"]["args"]:
|
||||
raw_result.update(
|
||||
{col_mapping[index]: out["server_command"]["args"][arg]}
|
||||
)
|
||||
|
||||
# Parse Client Command Arg
|
||||
out: dict[str, Any] = {
|
||||
"client_command": parse_client_command(command["client_command"])
|
||||
}
|
||||
parse_args = [
|
||||
"--dataset-name",
|
||||
"--random-input-len",
|
||||
"--random-output-len",
|
||||
"--request-rate",
|
||||
]
|
||||
col_mapping = ["dataset_name", "input_len", "output_len", "qps"]
|
||||
|
||||
for index, arg in enumerate(parse_args):
|
||||
if arg in out["client_command"]["args"]:
|
||||
raw_result.update(
|
||||
{col_mapping[index]: out["client_command"]["args"][arg]}
|
||||
)
|
||||
# Add Server, Client command
|
||||
raw_result.update(command)
|
||||
|
||||
# update the test name of this result
|
||||
raw_result.update({"test_name": test_file.stem})
|
||||
|
||||
# add the result to raw_result
|
||||
serving_results.append(raw_result)
|
||||
continue
|
||||
|
||||
elif "latency" in f.name:
|
||||
# this result is generated via `vllm bench latency` command
|
||||
# this result is generated via `benchmark_latency.py`
|
||||
|
||||
# attach the benchmarking command to raw_result
|
||||
try:
|
||||
@ -287,7 +148,7 @@ if __name__ == "__main__":
|
||||
continue
|
||||
|
||||
elif "throughput" in f.name:
|
||||
# this result is generated via `vllm bench throughput` command
|
||||
# this result is generated via `benchmark_throughput.py`
|
||||
|
||||
# attach the benchmarking command to raw_result
|
||||
try:
|
||||
@ -343,10 +204,7 @@ if __name__ == "__main__":
|
||||
columns=latency_column_mapping
|
||||
)
|
||||
if not serving_results.empty:
|
||||
valid_columns = [
|
||||
col for col in serving_column_mapping if col in serving_results.columns
|
||||
]
|
||||
serving_results = serving_results[valid_columns].rename(
|
||||
serving_results = serving_results[list(serving_column_mapping.keys())].rename(
|
||||
columns=serving_column_mapping
|
||||
)
|
||||
if not throughput_results.empty:
|
||||
@ -386,9 +244,7 @@ if __name__ == "__main__":
|
||||
)
|
||||
|
||||
# document the result
|
||||
md_file = "benchmark_results.md"
|
||||
json_file = "benchmark_results.json"
|
||||
with open(results_folder / md_file, "w") as f:
|
||||
with open(results_folder / "benchmark_results.md", "w") as f:
|
||||
results = read_markdown(
|
||||
"../.buildkite/nightly-benchmarks/"
|
||||
+ "performance-benchmarks-descriptions.md"
|
||||
@ -403,7 +259,7 @@ if __name__ == "__main__":
|
||||
f.write(results)
|
||||
|
||||
# document benchmarking results in json
|
||||
with open(results_folder / json_file, "w") as f:
|
||||
with open(results_folder / "benchmark_results.json", "w") as f:
|
||||
results = (
|
||||
latency_results.to_dict(orient="records")
|
||||
+ throughput_results.to_dict(orient="records")
|
||||
|
||||
@ -73,7 +73,7 @@ get_current_llm_serving_engine() {
|
||||
echo "Container: vllm"
|
||||
# move to a completely irrelevant directory, to avoid import vllm from current folder
|
||||
export CURRENT_LLM_SERVING_ENGINE=vllm
|
||||
|
||||
|
||||
return
|
||||
fi
|
||||
}
|
||||
@ -95,14 +95,12 @@ json2args() {
|
||||
}
|
||||
|
||||
kill_gpu_processes() {
|
||||
pkill -f '[p]ython'
|
||||
pkill -f '[p]ython3'
|
||||
pkill -f '[t]ritonserver'
|
||||
pkill -f '[p]t_main_thread'
|
||||
pkill -f '[t]ext-generation'
|
||||
pkill -f '[l]mdeploy'
|
||||
# vLLM now names the process with VLLM prefix after https://github.com/vllm-project/vllm/pull/21445
|
||||
pkill -f '[V]LLM'
|
||||
pkill -f python
|
||||
pkill -f python3
|
||||
pkill -f tritonserver
|
||||
pkill -f pt_main_thread
|
||||
pkill -f text-generation
|
||||
pkill -f lmdeploy
|
||||
|
||||
while [ "$(nvidia-smi --query-gpu=memory.used --format=csv,noheader,nounits | head -n 1)" -ge 1000 ]; do
|
||||
sleep 1
|
||||
@ -127,7 +125,7 @@ ensure_installed() {
|
||||
}
|
||||
|
||||
run_serving_tests() {
|
||||
# run serving tests using `vllm bench serve` command
|
||||
# run serving tests using `benchmark_serving.py`
|
||||
# $1: a json file specifying serving test cases
|
||||
|
||||
local serving_test_file
|
||||
@ -227,7 +225,7 @@ run_serving_tests() {
|
||||
|
||||
if [[ "$dataset_name" = "sharegpt" ]]; then
|
||||
|
||||
client_command="vllm bench serve \
|
||||
client_command="python3 benchmark_serving.py \
|
||||
--backend $backend \
|
||||
--tokenizer /tokenizer_cache \
|
||||
--model $model \
|
||||
@ -248,7 +246,7 @@ run_serving_tests() {
|
||||
sonnet_output_len=$(echo "$common_params" | jq -r '.sonnet_output_len')
|
||||
sonnet_prefix_len=$(echo "$common_params" | jq -r '.sonnet_prefix_len')
|
||||
|
||||
client_command="vllm bench serve \
|
||||
client_command="python3 benchmark_serving.py \
|
||||
--backend $backend \
|
||||
--tokenizer /tokenizer_cache \
|
||||
--model $model \
|
||||
@ -267,13 +265,13 @@ run_serving_tests() {
|
||||
$client_args"
|
||||
|
||||
else
|
||||
|
||||
|
||||
echo "The dataset name must be either 'sharegpt' or 'sonnet'. Got $dataset_name."
|
||||
exit 1
|
||||
|
||||
fi
|
||||
|
||||
|
||||
|
||||
|
||||
echo "Running test case $test_name with qps $qps"
|
||||
echo "Client command: $client_command"
|
||||
@ -304,7 +302,7 @@ run_serving_tests() {
|
||||
}
|
||||
|
||||
run_genai_perf_tests() {
|
||||
# run genai-perf tests
|
||||
# run genai-perf tests
|
||||
|
||||
# $1: a json file specifying genai-perf test cases
|
||||
local genai_perf_test_file
|
||||
@ -313,14 +311,14 @@ run_genai_perf_tests() {
|
||||
# Iterate over genai-perf tests
|
||||
jq -c '.[]' "$genai_perf_test_file" | while read -r params; do
|
||||
# get the test name, and append the GPU type back to it.
|
||||
test_name=$(echo "$params" | jq -r '.test_name')
|
||||
|
||||
test_name=$(echo "$params" | jq -r '.test_name')
|
||||
|
||||
# if TEST_SELECTOR is set, only run the test cases that match the selector
|
||||
if [[ -n "$TEST_SELECTOR" ]] && [[ ! "$test_name" =~ $TEST_SELECTOR ]]; then
|
||||
echo "Skip test case $test_name."
|
||||
continue
|
||||
fi
|
||||
|
||||
|
||||
# prepend the current serving engine to the test name
|
||||
test_name=${CURRENT_LLM_SERVING_ENGINE}_${test_name}
|
||||
|
||||
@ -371,10 +369,10 @@ run_genai_perf_tests() {
|
||||
qps=$num_prompts
|
||||
echo "now qps is $qps"
|
||||
fi
|
||||
|
||||
|
||||
new_test_name=$test_name"_qps_"$qps
|
||||
backend=$CURRENT_LLM_SERVING_ENGINE
|
||||
|
||||
|
||||
if [[ "$backend" == *"vllm"* ]]; then
|
||||
backend="vllm"
|
||||
fi
|
||||
@ -382,7 +380,7 @@ run_genai_perf_tests() {
|
||||
client_command="genai-perf profile \
|
||||
-m $model \
|
||||
--service-kind openai \
|
||||
--backend "$backend" \
|
||||
--backend vllm \
|
||||
--endpoint-type chat \
|
||||
--streaming \
|
||||
--url localhost:$port \
|
||||
@ -415,7 +413,7 @@ prepare_dataset() {
|
||||
do
|
||||
cat sonnet.txt >> sonnet_4x.txt
|
||||
done
|
||||
|
||||
|
||||
}
|
||||
|
||||
main() {
|
||||
|
||||
@ -33,7 +33,7 @@ check_gpus() {
|
||||
|
||||
check_cpus() {
|
||||
# check the number of CPUs and NUMA Node and GPU type.
|
||||
declare -g numa_count=$(lscpu | grep "NUMA node(s):" | awk '{print $3}')
|
||||
declare -g numa_count=$(python3 -c "from numa import info;numa_size = info.get_num_configured_nodes(); print(numa_size)")
|
||||
if [[ $numa_count -gt 0 ]]; then
|
||||
echo "NUMA found."
|
||||
echo $numa_count
|
||||
@ -126,8 +126,7 @@ kill_gpu_processes() {
|
||||
ps -aux
|
||||
lsof -t -i:8000 | xargs -r kill -9
|
||||
pgrep python3 | xargs -r kill -9
|
||||
# vLLM now names the process with VLLM prefix after https://github.com/vllm-project/vllm/pull/21445
|
||||
pgrep VLLM | xargs -r kill -9
|
||||
|
||||
|
||||
# wait until GPU memory usage smaller than 1GB
|
||||
if command -v nvidia-smi; then
|
||||
@ -165,7 +164,7 @@ upload_to_buildkite() {
|
||||
}
|
||||
|
||||
run_latency_tests() {
|
||||
# run latency tests using `vllm bench latency` command
|
||||
# run latency tests using `benchmark_latency.py`
|
||||
# $1: a json file specifying latency test cases
|
||||
|
||||
local latency_test_file
|
||||
@ -194,11 +193,9 @@ run_latency_tests() {
|
||||
|
||||
# check if there is enough GPU to run the test
|
||||
tp=$(echo "$latency_params" | jq -r '.tensor_parallel_size')
|
||||
if [ "$ON_CPU" == "1" ]; then
|
||||
pp=$(echo "$latency_params" | jq -r '.pipeline_parallel_size')
|
||||
world_size=$(($tp*$pp))
|
||||
if [[ $numa_count -lt $world_size && -z "${REMOTE_HOST}" ]]; then
|
||||
echo "Required world-size $world_size but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
if [ "$ON_CPU" == "1" ];then
|
||||
if [[ $numa_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
@ -208,7 +205,7 @@ run_latency_tests() {
|
||||
fi
|
||||
fi
|
||||
|
||||
latency_command=" $latency_envs vllm bench latency \
|
||||
latency_command=" $latency_envs python3 benchmark_latency.py \
|
||||
--output-json $RESULTS_FOLDER/${test_name}.json \
|
||||
$latency_args"
|
||||
|
||||
@ -234,7 +231,7 @@ run_latency_tests() {
|
||||
}
|
||||
|
||||
run_throughput_tests() {
|
||||
# run throughput tests using `vllm bench throughput`
|
||||
# run throughput tests using `benchmark_throughput.py`
|
||||
# $1: a json file specifying throughput test cases
|
||||
|
||||
local throughput_test_file
|
||||
@ -263,11 +260,9 @@ run_throughput_tests() {
|
||||
|
||||
# check if there is enough GPU to run the test
|
||||
tp=$(echo "$throughput_params" | jq -r '.tensor_parallel_size')
|
||||
if [ "$ON_CPU" == "1" ]; then
|
||||
pp=$(echo "$throughput_params" | jq -r '.pipeline_parallel_size')
|
||||
world_size=$(($tp*$pp))
|
||||
if [[ $numa_count -lt $world_size && -z "${REMOTE_HOST}" ]]; then
|
||||
echo "Required world-size $world_size but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
if [ "$ON_CPU" == "1" ];then
|
||||
if [[ $numa_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
@ -277,7 +272,7 @@ run_throughput_tests() {
|
||||
fi
|
||||
fi
|
||||
|
||||
throughput_command=" $throughput_envs vllm bench throughput \
|
||||
throughput_command=" $throughput_envs python3 benchmark_throughput.py \
|
||||
--output-json $RESULTS_FOLDER/${test_name}.json \
|
||||
$throughput_args"
|
||||
|
||||
@ -302,7 +297,7 @@ run_throughput_tests() {
|
||||
}
|
||||
|
||||
run_serving_tests() {
|
||||
# run serving tests using `vllm bench serve` command
|
||||
# run serving tests using `benchmark_serving.py`
|
||||
# $1: a json file specifying serving test cases
|
||||
|
||||
local serving_test_file
|
||||
@ -333,21 +328,12 @@ run_serving_tests() {
|
||||
qps_list=$(echo "$params" | jq -r '.qps_list')
|
||||
qps_list=$(echo "$qps_list" | jq -r '.[] | @sh')
|
||||
echo "Running over qps list $qps_list"
|
||||
max_concurrency_list=$(echo "$params" | jq -r '.max_concurrency_list')
|
||||
if [[ -z "$max_concurrency_list" || "$max_concurrency_list" == "null" ]]; then
|
||||
num_prompts=$(echo "$client_params" | jq -r '.num_prompts')
|
||||
max_concurrency_list="[$num_prompts]"
|
||||
fi
|
||||
max_concurrency_list=$(echo "$max_concurrency_list" | jq -r '.[] | @sh')
|
||||
echo "Running over max concurrency list $max_concurrency_list"
|
||||
|
||||
# check if there is enough resources to run the test
|
||||
tp=$(echo "$server_params" | jq -r '.tensor_parallel_size')
|
||||
if [ "$ON_CPU" == "1" ]; then
|
||||
pp=$(echo "$server_params" | jq -r '.pipeline_parallel_size')
|
||||
world_size=$(($tp*$pp))
|
||||
if [[ $numa_count -lt $world_size && -z "${REMOTE_HOST}" ]]; then
|
||||
echo "Required world-size $world_size but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
if [ "$ON_CPU" == "1" ];then
|
||||
if [[ $numa_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $numa_count NUMA nodes found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
@ -403,39 +389,35 @@ run_serving_tests() {
|
||||
echo "now qps is $qps"
|
||||
fi
|
||||
|
||||
# iterate over different max_concurrency
|
||||
for max_concurrency in $max_concurrency_list; do
|
||||
new_test_name=$test_name"_qps_"$qps"_concurrency_"$max_concurrency
|
||||
echo " new test name $new_test_name"
|
||||
# pass the tensor parallel size to the client so that it can be displayed
|
||||
# on the benchmark dashboard
|
||||
client_command="vllm bench serve \
|
||||
--save-result \
|
||||
--result-dir $RESULTS_FOLDER \
|
||||
--result-filename ${new_test_name}.json \
|
||||
--request-rate $qps \
|
||||
--max-concurrency $max_concurrency \
|
||||
--metadata "tensor_parallel_size=$tp" \
|
||||
$client_args $client_remote_args "
|
||||
new_test_name=$test_name"_qps_"$qps
|
||||
|
||||
echo "Running test case $test_name with qps $qps"
|
||||
echo "Client command: $client_command"
|
||||
# pass the tensor parallel size to the client so that it can be displayed
|
||||
# on the benchmark dashboard
|
||||
client_command="python3 benchmark_serving.py \
|
||||
--save-result \
|
||||
--result-dir $RESULTS_FOLDER \
|
||||
--result-filename ${new_test_name}.json \
|
||||
--request-rate $qps \
|
||||
--metadata "tensor_parallel_size=$tp" \
|
||||
$client_args $client_remote_args "
|
||||
|
||||
bash -c "$client_command"
|
||||
echo "Running test case $test_name with qps $qps"
|
||||
echo "Client command: $client_command"
|
||||
|
||||
# record the benchmarking commands
|
||||
jq_output=$(jq -n \
|
||||
--arg server "$server_command" \
|
||||
--arg client "$client_command" \
|
||||
--arg gpu "$gpu_type" \
|
||||
'{
|
||||
server_command: $server,
|
||||
client_command: $client,
|
||||
gpu_type: $gpu
|
||||
}')
|
||||
echo "$jq_output" >"$RESULTS_FOLDER/${new_test_name}.commands"
|
||||
bash -c "$client_command"
|
||||
|
||||
# record the benchmarking commands
|
||||
jq_output=$(jq -n \
|
||||
--arg server "$server_command" \
|
||||
--arg client "$client_command" \
|
||||
--arg gpu "$gpu_type" \
|
||||
'{
|
||||
server_command: $server,
|
||||
client_command: $client,
|
||||
gpu_type: $gpu
|
||||
}')
|
||||
echo "$jq_output" >"$RESULTS_FOLDER/${new_test_name}.commands"
|
||||
|
||||
done
|
||||
done
|
||||
|
||||
# clean up
|
||||
@ -465,7 +447,7 @@ main() {
|
||||
(which jq) || (apt-get update && apt-get -y install jq)
|
||||
(which lsof) || (apt-get update && apt-get install -y lsof)
|
||||
|
||||
# get the current IP address, required by `vllm bench serve` command
|
||||
# get the current IP address, required by benchmark_serving.py
|
||||
export VLLM_HOST_IP=$(hostname -I | awk '{print $1}')
|
||||
# turn of the reporting of the status of each request, to clean up the terminal output
|
||||
export VLLM_LOGGING_LEVEL="WARNING"
|
||||
|
||||
@ -11,7 +11,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
|
||||
@ -6,7 +6,7 @@
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"load_format": "dummy",
|
||||
"num_iters_warmup": 5,
|
||||
@ -20,7 +20,7 @@
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"load_format": "dummy",
|
||||
"num_iters_warmup": 5,
|
||||
|
||||
@ -35,7 +35,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
@ -88,7 +90,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
@ -141,7 +145,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
@ -191,7 +197,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
@ -243,7 +251,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
@ -295,7 +305,9 @@
|
||||
},
|
||||
"vllm_server_parameters": {
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"gpu_memory_utilization": 0.9,
|
||||
"num_scheduler_steps": 10,
|
||||
"max_num_seqs": 512,
|
||||
"dtype": "bfloat16"
|
||||
},
|
||||
|
||||
@ -1,202 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "serving_llama8B_tp1_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp2_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 2,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp4_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp1_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp2_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 2,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp4_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"num_prompts": 1000
|
||||
}
|
||||
}
|
||||
]
|
||||
@ -1,205 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "serving_llama8B_pp1_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"pipeline_parallel_size": 1,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_pp3_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"pipeline_parallel_size": 3,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp2pp3_sharegpt",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 2,
|
||||
"pipeline_parallel_size": 3,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_pp1_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"pipeline_parallel_size": 1,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_pp3_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"pipeline_parallel_size": 3,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp2pp3_random_128_128",
|
||||
"qps_list": ["inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200, 1000],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 2,
|
||||
"pipeline_parallel_size": 3,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
}
|
||||
]
|
||||
@ -2,112 +2,104 @@
|
||||
{
|
||||
"test_name": "serving_llama8B_tp1_sharegpt",
|
||||
"qps_list": [1, 4, 16, "inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"max_concurrency": 60,
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp2_sharegpt",
|
||||
"qps_list": [1, 4, 16, "inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 2,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"max_concurrency": 60,
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp4_sharegpt",
|
||||
"qps_list": [1, 4, 16, "inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
"block_size": 128,
|
||||
"trust_remote_code": "",
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"max_concurrency": 60,
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_tp4_random_1024_128",
|
||||
"qps_list": [1, 4, 16, "inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
@ -115,34 +107,32 @@
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 1024,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"max_concurrency": 100,
|
||||
"num_prompts": 100
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_pp6_random_1024_128",
|
||||
"qps_list": [1, 4, 16, "inf"],
|
||||
"max_concurrency_list": [12, 16, 24, 32, 64, 128, 200],
|
||||
"server_environment_variables": {
|
||||
"VLLM_RPC_TIMEOUT": 100000,
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_ENGINE_ITERATION_TIMEOUT_S": 120,
|
||||
"VLLM_CPU_SGL_KERNEL": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"pipeline_parallel_size": 6,
|
||||
"dtype": "bfloat16",
|
||||
"distributed_executor_backend": "mp",
|
||||
@ -150,18 +140,18 @@
|
||||
"trust_remote_code": "",
|
||||
"enable_chunked_prefill": "",
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"enforce_eager": "",
|
||||
"max_num_batched_tokens": 2048,
|
||||
"max_num_seqs": 256,
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 1024,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"max_concurrency": 100,
|
||||
"num_prompts": 100
|
||||
}
|
||||
}
|
||||
|
||||
@ -7,6 +7,7 @@
|
||||
"tensor_parallel_size": 1,
|
||||
"swap_space": 16,
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
@ -25,6 +26,7 @@
|
||||
"tensor_parallel_size": 4,
|
||||
"swap_space": 16,
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
@ -43,6 +45,7 @@
|
||||
"tensor_parallel_size": 2,
|
||||
"swap_space": 16,
|
||||
"disable_log_stats": "",
|
||||
"disable_log_requests": "",
|
||||
"load_format": "dummy"
|
||||
},
|
||||
"client_parameters": {
|
||||
@ -57,7 +60,8 @@
|
||||
"test_name": "serving_llama70B_tp4_sharegpt_specdecode",
|
||||
"qps_list": [2],
|
||||
"server_parameters": {
|
||||
"model": "meta-llama/Meta-Llama-3.1-70B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-70B-Instruct",
|
||||
"disable_log_requests": "",
|
||||
"tensor_parallel_size": 4,
|
||||
"swap_space": 16,
|
||||
"speculative_config": {
|
||||
|
||||
@ -6,7 +6,7 @@
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"load_format": "dummy",
|
||||
"dataset": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
@ -21,7 +21,7 @@
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"load_format": "dummy",
|
||||
"dataset": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
|
||||
@ -1,24 +1,5 @@
|
||||
steps:
|
||||
# aarch64 + CUDA builds. PyTorch 2.8 aarch64 + CUDA wheel is only available on CUDA 12.9
|
||||
- label: "Build arm64 wheel - CUDA 12.9"
|
||||
id: build-wheel-arm64-cuda-12-9
|
||||
agents:
|
||||
queue: arm64_cpu_queue_postmerge
|
||||
commands:
|
||||
# #NOTE: torch_cuda_arch_list is derived from upstream PyTorch build files here:
|
||||
# https://github.com/pytorch/pytorch/blob/main/.ci/aarch64_linux/aarch64_ci_build.sh#L7
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list='8.7 9.0 10.0+PTX 12.0' --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile ."
|
||||
- "mkdir artifacts"
|
||||
- "docker run --rm -v $(pwd)/artifacts:/artifacts_host vllm-ci:build-image bash -c 'cp -r dist /artifacts_host && chmod -R a+rw /artifacts_host'"
|
||||
- "bash .buildkite/scripts/upload-wheels.sh"
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- block: "Build CUDA 12.8 wheel"
|
||||
key: block-build-cu128-wheel
|
||||
|
||||
- label: "Build wheel - CUDA 12.8"
|
||||
depends_on: block-build-cu128-wheel
|
||||
id: build-wheel-cuda-12-8
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
@ -30,12 +11,7 @@ steps:
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- block: "Build CUDA 12.6 wheel"
|
||||
key: block-build-cu126-wheel
|
||||
depends_on: ~
|
||||
|
||||
- label: "Build wheel - CUDA 12.6"
|
||||
depends_on: block-build-cu126-wheel
|
||||
id: build-wheel-cuda-12-6
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
@ -47,63 +23,44 @@ steps:
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
# x86 + CUDA builds
|
||||
- label: "Build wheel - CUDA 12.9"
|
||||
depends_on: ~
|
||||
id: build-wheel-cuda-12-9
|
||||
# Note(simon): We can always build CUDA 11.8 wheel to ensure the build is working.
|
||||
# However, this block can be uncommented to save some compute hours.
|
||||
# - block: "Build CUDA 11.8 wheel"
|
||||
# key: block-build-cu118-wheel
|
||||
|
||||
- label: "Build wheel - CUDA 11.8"
|
||||
# depends_on: block-build-cu118-wheel
|
||||
id: build-wheel-cuda-11-8
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
commands:
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0+PTX' --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile ."
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=11.8.0 --build-arg torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0+PTX' --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile ."
|
||||
- "mkdir artifacts"
|
||||
- "docker run --rm -v $(pwd)/artifacts:/artifacts_host vllm-ci:build-image bash -c 'cp -r dist /artifacts_host && chmod -R a+rw /artifacts_host'"
|
||||
- "bash .buildkite/scripts/upload-wheels.sh"
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- label: "Build release image (x86)"
|
||||
- block: "Build release image"
|
||||
depends_on: ~
|
||||
id: build-release-image-x86
|
||||
key: block-release-image-build
|
||||
|
||||
- label: "Build release image"
|
||||
depends_on: block-release-image-build
|
||||
id: build-release-image
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
commands:
|
||||
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.8.1 --build-arg FLASHINFER_AOT_COMPILE=true --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)"
|
||||
# re-tag to default image tag and push, just in case arm64 build fails
|
||||
- "docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.8.1 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT --target vllm-openai --progress plain -f docker/Dockerfile ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT"
|
||||
|
||||
# PyTorch 2.8 aarch64 + CUDA wheel is only available on CUDA 12.9
|
||||
- label: "Build release image (arm64)"
|
||||
depends_on: ~
|
||||
id: build-release-image-arm64
|
||||
agents:
|
||||
queue: arm64_cpu_queue_postmerge
|
||||
commands:
|
||||
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list='8.7 9.0 10.0+PTX 12.0' --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)"
|
||||
|
||||
# Add job to create multi-arch manifest
|
||||
- label: "Create multi-arch manifest"
|
||||
depends_on:
|
||||
- build-release-image-x86
|
||||
- build-release-image-arm64
|
||||
id: create-multi-arch-manifest
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
commands:
|
||||
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
|
||||
- "docker manifest create public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-x86_64 public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-aarch64 --amend"
|
||||
- "docker manifest push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT"
|
||||
|
||||
- label: "Annotate release workflow"
|
||||
depends_on:
|
||||
- create-multi-arch-manifest
|
||||
- build-release-image
|
||||
- build-wheel-cuda-12-8
|
||||
- build-wheel-cuda-12-6
|
||||
- build-wheel-cuda-12-9
|
||||
- build-wheel-cuda-11-8
|
||||
id: annotate-release-workflow
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
|
||||
@ -107,8 +107,10 @@ fi
|
||||
|
||||
if [[ $commands == *" kernels/attention"* ]]; then
|
||||
commands="${commands} \
|
||||
--ignore=kernels/attention/test_attention_selector.py \
|
||||
--ignore=kernels/attention/stest_attention_selector.py \
|
||||
--ignore=kernels/attention/test_blocksparse_attention.py \
|
||||
--ignore=kernels/attention/test_encoder_decoder_attn.py \
|
||||
--ignore=kernels/attention/test_attention_selector.py \
|
||||
--ignore=kernels/attention/test_flash_attn.py \
|
||||
--ignore=kernels/attention/test_flashinfer.py \
|
||||
--ignore=kernels/attention/test_prefix_prefill.py \
|
||||
@ -121,6 +123,7 @@ fi
|
||||
if [[ $commands == *" kernels/quantization"* ]]; then
|
||||
commands="${commands} \
|
||||
--ignore=kernels/quantization/test_int8_quant.py \
|
||||
--ignore=kernels/quantization/test_aqlm.py \
|
||||
--ignore=kernels/quantization/test_machete_mm.py \
|
||||
--ignore=kernels/quantization/test_block_fp8.py \
|
||||
--ignore=kernels/quantization/test_block_int8.py \
|
||||
@ -164,6 +167,7 @@ if [[ $commands == *" entrypoints/llm "* ]]; then
|
||||
--ignore=entrypoints/llm/test_chat.py \
|
||||
--ignore=entrypoints/llm/test_accuracy.py \
|
||||
--ignore=entrypoints/llm/test_init.py \
|
||||
--ignore=entrypoints/llm/test_generate_multiple_loras.py \
|
||||
--ignore=entrypoints/llm/test_prompt_validation.py "}
|
||||
fi
|
||||
|
||||
|
||||
@ -6,16 +6,15 @@ set -ex
|
||||
|
||||
# allow to bind to different cores
|
||||
CORE_RANGE=${CORE_RANGE:-48-95}
|
||||
# used for TP/PP E2E test
|
||||
OMP_CORE_RANGE=${OMP_CORE_RANGE:-48-95}
|
||||
NUMA_NODE=${NUMA_NODE:-1}
|
||||
|
||||
export CMAKE_BUILD_PARALLEL_LEVEL=32
|
||||
|
||||
# Setup cleanup
|
||||
remove_docker_container() {
|
||||
set -e;
|
||||
docker rm -f cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"-avx2 || true;
|
||||
remove_docker_container() {
|
||||
set -e;
|
||||
docker rm -f cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"-avx2 || true;
|
||||
}
|
||||
trap remove_docker_container EXIT
|
||||
remove_docker_container
|
||||
@ -25,8 +24,8 @@ numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --tag cpu-test-"$NUMA_NODE
|
||||
numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --build-arg VLLM_CPU_DISABLE_AVX512="true" --tag cpu-test-"$NUMA_NODE"-avx2 --target vllm-test -f docker/Dockerfile.cpu .
|
||||
|
||||
# Run the image, setting --shm-size=4g for tensor parallel.
|
||||
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=16 --env VLLM_CPU_CI_ENV=1 -e E2E_OMP_THREADS="$OMP_CORE_RANGE" --shm-size=4g --name cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"
|
||||
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=16 --env VLLM_CPU_CI_ENV=1 -e E2E_OMP_THREADS="$OMP_CORE_RANGE" --shm-size=4g --name cpu-test-"$NUMA_NODE"-avx2 cpu-test-"$NUMA_NODE"-avx2
|
||||
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --env VLLM_CPU_CI_ENV=1 --shm-size=4g --name cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"
|
||||
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --env VLLM_CPU_CI_ENV=1 --shm-size=4g --name cpu-test-"$NUMA_NODE"-avx2 cpu-test-"$NUMA_NODE"-avx2
|
||||
|
||||
function cpu_tests() {
|
||||
set -e
|
||||
@ -46,26 +45,15 @@ function cpu_tests() {
|
||||
set -e
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m"
|
||||
|
||||
# Run kernel tests
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
pytest -x -v -s tests/kernels/test_onednn.py"
|
||||
|
||||
# Run basic model test
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
# Note: disable until supports V1
|
||||
# pytest -x -v -s tests/kernels/attention/test_cache.py -m cpu_model
|
||||
# pytest -x -v -s tests/kernels/attention/test_mla_decode_cpu.py -m cpu_model
|
||||
|
||||
# Note: disable Bart until supports V1
|
||||
pytest -x -v -s tests/models/language/generation -m cpu_model \
|
||||
--ignore=tests/models/language/generation/test_bart.py
|
||||
VLLM_CPU_SGL_KERNEL=1 pytest -x -v -s tests/models/language/generation -m cpu_model \
|
||||
--ignore=tests/models/language/generation/test_bart.py
|
||||
|
||||
pytest -x -v -s tests/models/language/pooling -m cpu_model
|
||||
pytest -x -v -s tests/models/multimodal/generation \
|
||||
pytest -v -s tests/kernels/attention/test_cache.py -m cpu_model
|
||||
pytest -v -s tests/kernels/attention/test_mla_decode_cpu.py -m cpu_model
|
||||
pytest -v -s tests/models/language/generation -m cpu_model
|
||||
VLLM_CPU_SGL_KERNEL=1 pytest -v -s tests/models/language/generation -m cpu_model
|
||||
pytest -v -s tests/models/language/pooling -m cpu_model
|
||||
pytest -v -s tests/models/multimodal/generation \
|
||||
--ignore=tests/models/multimodal/generation/test_mllama.py \
|
||||
--ignore=tests/models/multimodal/generation/test_pixtral.py \
|
||||
-m cpu_model"
|
||||
@ -73,51 +61,42 @@ function cpu_tests() {
|
||||
# Run compressed-tensor test
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
pytest -x -s -v \
|
||||
tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_logprobs[False-10-32-neuralmagic/Llama-3.2-1B-quantized.w8a8]"
|
||||
pytest -s -v \
|
||||
tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_static_setup \
|
||||
tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_dynamic_per_token"
|
||||
|
||||
# Note: disable it until supports V1
|
||||
# Run AWQ test
|
||||
# docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
# set -e
|
||||
# VLLM_USE_V1=0 pytest -x -s -v \
|
||||
# tests/quantization/test_ipex_quant.py"
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
VLLM_USE_V1=0 pytest -s -v \
|
||||
tests/quantization/test_ipex_quant.py"
|
||||
|
||||
# Run chunked-prefill and prefix-cache test
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
pytest -s -v -k cpu_model \
|
||||
tests/basic_correctness/test_chunked_prefill.py"
|
||||
|
||||
# online serving
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
python3 -m vllm.entrypoints.openai.api_server --model facebook/opt-125m --dtype half &
|
||||
timeout 600 bash -c 'until curl localhost:8000/v1/models; do sleep 1; done' || exit 1
|
||||
VLLM_CPU_CI_ENV=0 python3 benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--dataset-name random \
|
||||
--model facebook/opt-125m \
|
||||
--num-prompts 20 \
|
||||
--endpoint /v1/completions \
|
||||
--tokenizer facebook/opt-125m"
|
||||
|
||||
# Run multi-lora tests
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c "
|
||||
set -e
|
||||
pytest -x -s -v \
|
||||
pytest -s -v \
|
||||
tests/lora/test_qwen2vl.py"
|
||||
|
||||
# online serving: tp+pp
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c '
|
||||
set -e
|
||||
VLLM_CPU_OMP_THREADS_BIND=$E2E_OMP_THREADS VLLM_CPU_SGL_KERNEL=1 vllm serve meta-llama/Llama-3.2-3B-Instruct -tp=2 -pp=2 &
|
||||
server_pid=$!
|
||||
timeout 600 bash -c "until curl localhost:8000/v1/models; do sleep 1; done" || exit 1
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--dataset-name random \
|
||||
--model meta-llama/Llama-3.2-3B-Instruct \
|
||||
--num-prompts 20 \
|
||||
--endpoint /v1/completions
|
||||
kill -s SIGTERM $server_pid &'
|
||||
|
||||
# online serving: tp+dp
|
||||
docker exec cpu-test-"$NUMA_NODE" bash -c '
|
||||
set -e
|
||||
VLLM_CPU_OMP_THREADS_BIND=$E2E_OMP_THREADS VLLM_CPU_SGL_KERNEL=1 vllm serve meta-llama/Llama-3.2-3B-Instruct -tp=2 -dp=2 &
|
||||
server_pid=$!
|
||||
timeout 600 bash -c "until curl localhost:8000/v1/models; do sleep 1; done" || exit 1
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--dataset-name random \
|
||||
--model meta-llama/Llama-3.2-3B-Instruct \
|
||||
--num-prompts 20 \
|
||||
--endpoint /v1/completions
|
||||
kill -s SIGTERM $server_pid &'
|
||||
}
|
||||
|
||||
# All of CPU tests are expected to be finished less than 40 mins.
|
||||
export -f cpu_tests
|
||||
timeout 2h bash -c "cpu_tests $CORE_RANGE $NUMA_NODE"
|
||||
timeout 1.5h bash -c "cpu_tests $CORE_RANGE $NUMA_NODE"
|
||||
|
||||
@ -16,7 +16,8 @@ DOCKER_BUILDKIT=1 docker build . \
|
||||
--build-arg max_jobs=66 \
|
||||
--build-arg nvcc_threads=2 \
|
||||
--build-arg RUN_WHEEL_CHECK=false \
|
||||
--build-arg torch_cuda_arch_list="9.0+PTX"
|
||||
--build-arg torch_cuda_arch_list="9.0+PTX" \
|
||||
--build-arg vllm_fa_cmake_gpu_arches="90-real"
|
||||
|
||||
# Setup cleanup
|
||||
remove_docker_container() { docker rm -f gh200-test || true; }
|
||||
|
||||
@ -6,17 +6,19 @@ set -exuo pipefail
|
||||
|
||||
# Try building the docker image
|
||||
cat <<EOF | docker build -t hpu-plugin-v1-test-env -f - .
|
||||
FROM gaudi-base-image:latest
|
||||
FROM 1.22-413-pt2.7.1:latest
|
||||
|
||||
COPY ./ /workspace/vllm
|
||||
|
||||
WORKDIR /workspace/vllm
|
||||
|
||||
RUN pip install -v -r requirements/hpu.txt
|
||||
RUN pip install git+https://github.com/vllm-project/vllm-gaudi.git
|
||||
|
||||
ENV no_proxy=localhost,127.0.0.1
|
||||
ENV PT_HPU_ENABLE_LAZY_COLLECTIVES=true
|
||||
|
||||
RUN VLLM_TARGET_DEVICE=empty pip install .
|
||||
RUN pip install git+https://github.com/vllm-project/vllm-gaudi.git
|
||||
RUN VLLM_TARGET_DEVICE=hpu python3 setup.py install
|
||||
|
||||
# install development dependencies (for testing)
|
||||
RUN python3 -m pip install -e tests/vllm_test_utils
|
||||
|
||||
@ -1,167 +0,0 @@
|
||||
#!/bin/bash
|
||||
|
||||
set -xu
|
||||
|
||||
|
||||
remove_docker_container() {
|
||||
docker rm -f tpu-test || true;
|
||||
}
|
||||
|
||||
trap remove_docker_container EXIT
|
||||
|
||||
# Remove the container that might not be cleaned up in the previous run.
|
||||
remove_docker_container
|
||||
|
||||
# Build the docker image.
|
||||
docker build -f docker/Dockerfile.tpu -t vllm-tpu .
|
||||
|
||||
# Set up cleanup.
|
||||
cleanup_docker() {
|
||||
# Get Docker's root directory
|
||||
docker_root=$(docker info -f '{{.DockerRootDir}}')
|
||||
if [ -z "$docker_root" ]; then
|
||||
echo "Failed to determine Docker root directory."
|
||||
exit 1
|
||||
fi
|
||||
echo "Docker root directory: $docker_root"
|
||||
# Check disk usage of the filesystem where Docker's root directory is located
|
||||
disk_usage=$(df "$docker_root" | tail -1 | awk '{print $5}' | sed 's/%//')
|
||||
# Define the threshold
|
||||
threshold=70
|
||||
if [ "$disk_usage" -gt "$threshold" ]; then
|
||||
echo "Disk usage is above $threshold%. Cleaning up Docker images and volumes..."
|
||||
# Remove dangling images (those that are not tagged and not used by any container)
|
||||
docker image prune -f
|
||||
# Remove unused volumes / force the system prune for old images as well.
|
||||
docker volume prune -f && docker system prune --force --filter "until=72h" --all
|
||||
echo "Docker images and volumes cleanup completed."
|
||||
else
|
||||
echo "Disk usage is below $threshold%. No cleanup needed."
|
||||
fi
|
||||
}
|
||||
cleanup_docker
|
||||
|
||||
# For HF_TOKEN.
|
||||
source /etc/environment
|
||||
|
||||
docker run --privileged --net host --shm-size=16G -it \
|
||||
-e "HF_TOKEN=$HF_TOKEN" --name tpu-test \
|
||||
vllm-tpu /bin/bash -c '
|
||||
set -e # Exit immediately if a command exits with a non-zero status.
|
||||
set -u # Treat unset variables as an error.
|
||||
|
||||
echo "--- Starting script inside Docker container ---"
|
||||
|
||||
# Create results directory
|
||||
RESULTS_DIR=$(mktemp -d)
|
||||
# If mktemp fails, set -e will cause the script to exit.
|
||||
echo "Results will be stored in: $RESULTS_DIR"
|
||||
|
||||
# Install dependencies
|
||||
echo "--- Installing Python dependencies ---"
|
||||
python3 -m pip install --progress-bar off git+https://github.com/thuml/depyf.git \
|
||||
&& python3 -m pip install --progress-bar off pytest pytest-asyncio tpu-info \
|
||||
&& python3 -m pip install --progress-bar off "lm-eval @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d" \
|
||||
&& python3 -m pip install --progress-bar off hf-transfer
|
||||
echo "--- Python dependencies installed ---"
|
||||
export VLLM_USE_V1=1
|
||||
export VLLM_XLA_CHECK_RECOMPILATION=1
|
||||
export VLLM_XLA_CACHE_PATH=
|
||||
echo "Using VLLM V1"
|
||||
|
||||
echo "--- Hardware Information ---"
|
||||
# tpu-info
|
||||
echo "--- Starting Tests ---"
|
||||
set +e
|
||||
overall_script_exit_code=0
|
||||
|
||||
# --- Test Definitions ---
|
||||
# If a test fails, this function will print logs and will not cause the main script to exit.
|
||||
run_test() {
|
||||
local test_num=$1
|
||||
local test_name=$2
|
||||
local test_command=$3
|
||||
local log_file="$RESULTS_DIR/test_${test_num}.log"
|
||||
local actual_exit_code
|
||||
|
||||
echo "--- TEST_$test_num: Running $test_name ---"
|
||||
|
||||
# Execute the test command.
|
||||
eval "$test_command" > >(tee -a "$log_file") 2> >(tee -a "$log_file" >&2)
|
||||
actual_exit_code=$?
|
||||
|
||||
echo "TEST_${test_num}_COMMAND_EXIT_CODE: $actual_exit_code" # This goes to main log
|
||||
echo "TEST_${test_num}_COMMAND_EXIT_CODE: $actual_exit_code" >> "$log_file" # Also to per-test log
|
||||
|
||||
if [ "$actual_exit_code" -ne 0 ]; then
|
||||
echo "TEST_$test_num ($test_name) FAILED with exit code $actual_exit_code." >&2
|
||||
echo "--- Log for failed TEST_$test_num ($test_name) ---" >&2
|
||||
if [ -f "$log_file" ]; then
|
||||
cat "$log_file" >&2
|
||||
else
|
||||
echo "Log file $log_file not found for TEST_$test_num ($test_name)." >&2
|
||||
fi
|
||||
echo "--- End of log for TEST_$test_num ($test_name) ---" >&2
|
||||
return "$actual_exit_code" # Return the failure code
|
||||
else
|
||||
echo "TEST_$test_num ($test_name) PASSED."
|
||||
return 0 # Return success
|
||||
fi
|
||||
}
|
||||
|
||||
# Helper function to call run_test and update the overall script exit code
|
||||
run_and_track_test() {
|
||||
local test_num_arg="$1"
|
||||
local test_name_arg="$2"
|
||||
local test_command_arg="$3"
|
||||
|
||||
# Run the test
|
||||
run_test "$test_num_arg" "$test_name_arg" "$test_command_arg"
|
||||
local test_specific_exit_code=$?
|
||||
|
||||
# If the test failed, set the overall script exit code to 1
|
||||
if [ "$test_specific_exit_code" -ne 0 ]; then
|
||||
# No need for extra echo here, run_test already logged the failure.
|
||||
overall_script_exit_code=1
|
||||
fi
|
||||
}
|
||||
|
||||
# --- Actual Test Execution ---
|
||||
run_and_track_test 1 "test_struct_output_generate.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/entrypoints/llm/test_struct_output_generate.py -k \"not test_structured_output_with_reasoning_matrices\""
|
||||
run_and_track_test 2 "test_moe_pallas.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/tpu/test_moe_pallas.py"
|
||||
run_and_track_test 3 "test_lora.py" \
|
||||
"VLLM_XLA_CHECK_RECOMPILATION=0 python3 -m pytest -s -v /workspace/vllm/tests/tpu/lora/test_lora.py"
|
||||
run_and_track_test 4 "test_tpu_qkv_linear.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_tpu_qkv_linear.py"
|
||||
run_and_track_test 5 "test_spmd_model_weight_loading.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_spmd_model_weight_loading.py"
|
||||
run_and_track_test 6 "test_kv_cache_update_kernel.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_kv_cache_update_kernel.py"
|
||||
run_and_track_test 7 "test_tpu_int8.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_tpu_int8.py"
|
||||
|
||||
# After all tests have been attempted, exit with the overall status.
|
||||
if [ "$overall_script_exit_code" -ne 0 ]; then
|
||||
echo "--- One or more tests FAILED. Overall script exiting with failure code 1. ---"
|
||||
else
|
||||
echo "--- All tests have completed and PASSED. Overall script exiting with success code 0. ---"
|
||||
fi
|
||||
exit "$overall_script_exit_code"
|
||||
' # IMPORTANT: This is the closing single quote for the bash -c "..." command. Ensure it is present and correct.
|
||||
|
||||
# Capture the exit code of the docker run command
|
||||
DOCKER_RUN_EXIT_CODE=$?
|
||||
|
||||
# The trap will run for cleanup.
|
||||
# Exit the main script with the Docker run command's exit code.
|
||||
if [ "$DOCKER_RUN_EXIT_CODE" -ne 0 ]; then
|
||||
echo "Docker run command failed with exit code $DOCKER_RUN_EXIT_CODE."
|
||||
exit "$DOCKER_RUN_EXIT_CODE"
|
||||
else
|
||||
echo "Docker run command completed successfully."
|
||||
exit 0
|
||||
fi
|
||||
# TODO: This test fails because it uses RANDOM_SEED sampling
|
||||
# pytest -v -s /workspace/vllm/tests/tpu/test_custom_dispatcher.py \
|
||||
@ -5,6 +5,7 @@ set -xu
|
||||
|
||||
remove_docker_container() {
|
||||
docker rm -f tpu-test || true;
|
||||
docker rm -f vllm-tpu || true;
|
||||
}
|
||||
|
||||
trap remove_docker_container EXIT
|
||||
@ -61,8 +62,7 @@ echo "Results will be stored in: $RESULTS_DIR"
|
||||
echo "--- Installing Python dependencies ---"
|
||||
python3 -m pip install --progress-bar off git+https://github.com/thuml/depyf.git \
|
||||
&& python3 -m pip install --progress-bar off pytest pytest-asyncio tpu-info \
|
||||
&& python3 -m pip install --progress-bar off "lm-eval @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d" \
|
||||
&& python3 -m pip install --progress-bar off hf-transfer
|
||||
&& python3 -m pip install --progress-bar off lm_eval[api]==0.4.4
|
||||
echo "--- Python dependencies installed ---"
|
||||
export VLLM_USE_V1=1
|
||||
export VLLM_XLA_CHECK_RECOMPILATION=1
|
||||
@ -70,7 +70,7 @@ export VLLM_XLA_CACHE_PATH=
|
||||
echo "Using VLLM V1"
|
||||
|
||||
echo "--- Hardware Information ---"
|
||||
# tpu-info
|
||||
tpu-info
|
||||
echo "--- Starting Tests ---"
|
||||
set +e
|
||||
overall_script_exit_code=0
|
||||
@ -149,6 +149,18 @@ run_and_track_test 9 "test_multimodal.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_multimodal.py"
|
||||
run_and_track_test 10 "test_pallas.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_pallas.py"
|
||||
run_and_track_test 11 "test_struct_output_generate.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/entrypoints/llm/test_struct_output_generate.py -k \"not test_structured_output_with_reasoning_matrices\""
|
||||
run_and_track_test 12 "test_moe_pallas.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/tpu/test_moe_pallas.py"
|
||||
run_and_track_test 13 "test_lora.py" \
|
||||
"VLLM_XLA_CHECK_RECOMPILATION=0 python3 -m pytest -s -v /workspace/vllm/tests/tpu/lora/test_lora.py"
|
||||
run_and_track_test 14 "test_tpu_qkv_linear.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_tpu_qkv_linear.py"
|
||||
run_and_track_test 15 "test_spmd_model_weight_loading.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_spmd_model_weight_loading.py"
|
||||
run_and_track_test 16 "test_kv_cache_update_kernel.py" \
|
||||
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_kv_cache_update_kernel.py"
|
||||
|
||||
# After all tests have been attempted, exit with the overall status.
|
||||
if [ "$overall_script_exit_code" -ne 0 ]; then
|
||||
|
||||
@ -11,8 +11,8 @@ container_name="xpu_${BUILDKITE_COMMIT}_$(tr -dc A-Za-z0-9 < /dev/urandom | head
|
||||
docker build -t ${image_name} -f docker/Dockerfile.xpu .
|
||||
|
||||
# Setup cleanup
|
||||
remove_docker_container() {
|
||||
docker rm -f "${container_name}" || true;
|
||||
remove_docker_container() {
|
||||
docker rm -f "${container_name}" || true;
|
||||
docker image rm -f "${image_name}" || true;
|
||||
docker system prune -f || true;
|
||||
}
|
||||
@ -23,27 +23,10 @@ docker run \
|
||||
--device /dev/dri \
|
||||
-v /dev/dri/by-path:/dev/dri/by-path \
|
||||
--entrypoint="" \
|
||||
-e "HF_TOKEN=${HF_TOKEN}" \
|
||||
-e "ZE_AFFINITY_MASK=${ZE_AFFINITY_MASK}" \
|
||||
--name "${container_name}" \
|
||||
"${image_name}" \
|
||||
bash -c '
|
||||
set -e
|
||||
echo $ZE_AFFINITY_MASK
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 -O3 -O.cudagraph_mode=NONE
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend ray
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend mp
|
||||
VLLM_ATTENTION_BACKEND=TRITON_ATTN_VLLM_V1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
|
||||
cd tests
|
||||
pytest -v -s v1/core
|
||||
pytest -v -s v1/engine
|
||||
pytest -v -s v1/sample --ignore=v1/sample/test_logprobs.py --ignore=v1/sample/test_logprobs_e2e.py
|
||||
pytest -v -s v1/worker --ignore=v1/worker/test_gpu_model_runner.py
|
||||
pytest -v -s v1/structured_output
|
||||
pytest -v -s v1/spec_decode --ignore=v1/spec_decode/test_max_len.py --ignore=v1/spec_decode/test_eagle.py --ignore=v1/spec_decode/test_tree_attention.py
|
||||
pytest -v -s v1/kv_connector/unit --ignore=v1/kv_connector/unit/test_multi_connector.py --ignore=v1/kv_connector/unit/test_nixl_connector.py --ignore=v1/kv_connector/unit/test_shared_storage_connector.py
|
||||
pytest -v -s v1/test_serial_utils.py
|
||||
pytest -v -s v1/test_utils.py
|
||||
pytest -v -s v1/test_metrics_reader.py
|
||||
sh -c '
|
||||
VLLM_USE_V1=0 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m
|
||||
VLLM_USE_V1=0 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m -tp 2
|
||||
VLLM_USE_V1=1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
|
||||
'
|
||||
|
||||
@ -11,10 +11,10 @@ cd "$(dirname "${BASH_SOURCE[0]}")/../.."
|
||||
(which wget && which curl) || (apt-get update && apt-get install -y wget curl)
|
||||
|
||||
# run python-based benchmarks and upload the result to buildkite
|
||||
vllm bench latency --output-json latency_results.json 2>&1 | tee benchmark_latency.txt
|
||||
python3 benchmarks/benchmark_latency.py --output-json latency_results.json 2>&1 | tee benchmark_latency.txt
|
||||
bench_latency_exit_code=$?
|
||||
|
||||
vllm bench throughput --input-len 256 --output-len 256 --output-json throughput_results.json 2>&1 | tee benchmark_throughput.txt
|
||||
python3 benchmarks/benchmark_throughput.py --input-len 256 --output-len 256 --output-json throughput_results.json 2>&1 | tee benchmark_throughput.txt
|
||||
bench_throughput_exit_code=$?
|
||||
|
||||
# run server-based benchmarks and upload the result to buildkite
|
||||
@ -24,7 +24,7 @@ wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/r
|
||||
|
||||
# wait for server to start, timeout after 600 seconds
|
||||
timeout 600 bash -c 'until curl localhost:8000/v1/models; do sleep 1; done' || exit 1
|
||||
vllm bench serve \
|
||||
python3 benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path ./ShareGPT_V3_unfiltered_cleaned_split.json \
|
||||
|
||||
@ -17,7 +17,7 @@ if [ "$disk_usage" -gt "$threshold" ]; then
|
||||
# Remove dangling images (those that are not tagged and not used by any container)
|
||||
docker image prune -f
|
||||
# Remove unused volumes / force the system prune for old images as well.
|
||||
docker volume prune -f && docker system prune --force --filter "until=24h" --all
|
||||
docker volume prune -f && docker system prune --force --filter "until=72h" --all
|
||||
echo "Docker images and volumes cleanup completed."
|
||||
else
|
||||
echo "Disk usage is below $threshold%. No cleanup needed."
|
||||
|
||||
@ -1,6 +1,6 @@
|
||||
# Environment config
|
||||
TEST_NAME=llama8b
|
||||
CONTAINER_NAME=tpu-test
|
||||
CONTAINER_NAME=vllm-tpu
|
||||
|
||||
# vllm config
|
||||
MODEL=meta-llama/Llama-3.1-8B-Instruct
|
||||
|
||||
@ -12,6 +12,8 @@ source /etc/environment
|
||||
source $ENV_FILE
|
||||
|
||||
remove_docker_container() {
|
||||
docker rm -f tpu-test || true;
|
||||
docker rm -f vllm-tpu || true;
|
||||
docker rm -f $CONTAINER_NAME || true;
|
||||
}
|
||||
|
||||
@ -20,6 +22,16 @@ trap remove_docker_container EXIT
|
||||
# Remove the container that might not be cleaned up in the previous run.
|
||||
remove_docker_container
|
||||
|
||||
# Build docker image.
|
||||
# TODO: build the image outside the script and share the image with other
|
||||
# tpu test if building time is too long.
|
||||
DOCKER_BUILDKIT=1 docker build \
|
||||
--build-arg max_jobs=16 \
|
||||
--build-arg USE_SCCACHE=1 \
|
||||
--build-arg GIT_REPO_CHECK=0 \
|
||||
--tag vllm/vllm-tpu-bm \
|
||||
--progress plain -f docker/Dockerfile.tpu .
|
||||
|
||||
LOG_ROOT=$(mktemp -d)
|
||||
# If mktemp fails, set -e will cause the script to exit.
|
||||
echo "Results will be stored in: $LOG_ROOT"
|
||||
|
||||
@ -1,6 +1,6 @@
|
||||
# Environment config
|
||||
TEST_NAME=llama8bw8a8
|
||||
CONTAINER_NAME=tpu-test
|
||||
CONTAINER_NAME=vllm-tpu
|
||||
|
||||
# vllm config
|
||||
MODEL=RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8
|
||||
|
||||
@ -44,6 +44,7 @@ echo
|
||||
|
||||
VLLM_USE_V1=1 vllm serve $MODEL \
|
||||
--seed 42 \
|
||||
--disable-log-requests \
|
||||
--max-num-seqs $MAX_NUM_SEQS \
|
||||
--max-num-batched-tokens $MAX_NUM_BATCHED_TOKENS \
|
||||
--tensor-parallel-size $TENSOR_PARALLEL_SIZE \
|
||||
@ -76,7 +77,7 @@ done
|
||||
echo "run benchmark test..."
|
||||
echo "logging to $BM_LOG"
|
||||
echo
|
||||
vllm bench serve \
|
||||
python benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $MODEL \
|
||||
--dataset-name sonnet \
|
||||
|
||||
@ -14,19 +14,8 @@ fi
|
||||
# Get the single wheel file
|
||||
wheel="${wheel_files[0]}"
|
||||
|
||||
# Detect architecture and rename 'linux' to appropriate manylinux version
|
||||
arch=$(uname -m)
|
||||
if [[ $arch == "x86_64" ]]; then
|
||||
manylinux_version="manylinux1"
|
||||
elif [[ $arch == "aarch64" ]]; then
|
||||
manylinux_version="manylinux2014"
|
||||
else
|
||||
echo "Warning: Unknown architecture $arch, using manylinux1 as default"
|
||||
manylinux_version="manylinux1"
|
||||
fi
|
||||
|
||||
# Rename 'linux' to the appropriate manylinux version in the wheel filename
|
||||
new_wheel="${wheel/linux/$manylinux_version}"
|
||||
# Rename 'linux' to 'manylinux1' in the wheel filename
|
||||
new_wheel="${wheel/linux/manylinux1}"
|
||||
mv -- "$wheel" "$new_wheel"
|
||||
wheel="$new_wheel"
|
||||
|
||||
@ -58,15 +47,14 @@ python3 .buildkite/generate_index.py --wheel "$normal_wheel"
|
||||
aws s3 cp "$wheel" "s3://vllm-wheels/$BUILDKITE_COMMIT/"
|
||||
aws s3 cp "$normal_wheel" "s3://vllm-wheels/$BUILDKITE_COMMIT/"
|
||||
|
||||
if [[ $normal_wheel == *"cu126"* ]]; then
|
||||
if [[ $normal_wheel == *"cu118"* ]]; then
|
||||
# if $normal_wheel matches cu118, do not upload the index.html
|
||||
echo "Skipping index files for cu118 wheels"
|
||||
elif [[ $normal_wheel == *"cu126"* ]]; then
|
||||
# if $normal_wheel matches cu126, do not upload the index.html
|
||||
echo "Skipping index files for cu126 wheels"
|
||||
elif [[ $normal_wheel == *"cu128"* ]]; then
|
||||
# if $normal_wheel matches cu128, do not upload the index.html
|
||||
echo "Skipping index files for cu128 wheels"
|
||||
else
|
||||
# only upload index.html for cu129 wheels (default wheels) as it
|
||||
# is available on both x86 and arm64
|
||||
# only upload index.html for cu128 wheels (default wheels)
|
||||
aws s3 cp index.html "s3://vllm-wheels/$BUILDKITE_COMMIT/vllm/index.html"
|
||||
aws s3 cp "s3://vllm-wheels/nightly/index.html" "s3://vllm-wheels/$BUILDKITE_COMMIT/index.html"
|
||||
fi
|
||||
@ -75,15 +63,14 @@ fi
|
||||
aws s3 cp "$wheel" "s3://vllm-wheels/nightly/"
|
||||
aws s3 cp "$normal_wheel" "s3://vllm-wheels/nightly/"
|
||||
|
||||
if [[ $normal_wheel == *"cu126"* ]]; then
|
||||
if [[ $normal_wheel == *"cu118"* ]]; then
|
||||
# if $normal_wheel matches cu118, do not upload the index.html
|
||||
echo "Skipping index files for cu118 wheels"
|
||||
elif [[ $normal_wheel == *"cu126"* ]]; then
|
||||
# if $normal_wheel matches cu126, do not upload the index.html
|
||||
echo "Skipping index files for cu126 wheels"
|
||||
elif [[ $normal_wheel == *"cu128"* ]]; then
|
||||
# if $normal_wheel matches cu128, do not upload the index.html
|
||||
echo "Skipping index files for cu128 wheels"
|
||||
else
|
||||
# only upload index.html for cu129 wheels (default wheels) as it
|
||||
# is available on both x86 and arm64
|
||||
# only upload index.html for cu128 wheels (default wheels)
|
||||
aws s3 cp index.html "s3://vllm-wheels/nightly/vllm/index.html"
|
||||
fi
|
||||
|
||||
|
||||
@ -31,6 +31,16 @@
|
||||
steps:
|
||||
##### fast check tests #####
|
||||
|
||||
- label: Documentation Build # 2min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
working_dir: "/vllm-workspace/test_docs"
|
||||
fast_check: true
|
||||
no_gpu: True
|
||||
commands:
|
||||
- pip install -r ../requirements/docs.txt
|
||||
# TODO: add `--strict` once warnings in docstrings are fixed
|
||||
- mkdocs build
|
||||
|
||||
- label: Pytorch Nightly Dependency Override Check # 2min
|
||||
# if this test fails, it means the nightly torch version is not compatible with some
|
||||
# of the dependencies. Please check the error message and add the package to whitelist
|
||||
@ -47,20 +57,20 @@ steps:
|
||||
- vllm/
|
||||
- tests/mq_llm_engine
|
||||
- tests/async_engine
|
||||
- tests/test_inputs.py
|
||||
- tests/test_outputs.py
|
||||
- tests/test_inputs
|
||||
- tests/multimodal
|
||||
- tests/utils_
|
||||
- tests/test_utils
|
||||
- tests/worker
|
||||
- tests/standalone_tests/lazy_imports.py
|
||||
commands:
|
||||
- python3 standalone_tests/lazy_imports.py
|
||||
- pytest -v -s mq_llm_engine # MQLLMEngine
|
||||
- pytest -v -s async_engine # AsyncLLMEngine
|
||||
- NUM_SCHEDULER_STEPS=4 pytest -v -s async_engine/test_async_llm_engine.py
|
||||
- pytest -v -s test_inputs.py
|
||||
- pytest -v -s test_outputs.py
|
||||
- pytest -v -s multimodal
|
||||
- pytest -v -s utils_ # Utils
|
||||
- pytest -v -s test_utils.py # Utils
|
||||
- pytest -v -s worker # Worker
|
||||
|
||||
- label: Python-only Installation Test
|
||||
@ -72,7 +82,7 @@ steps:
|
||||
- bash standalone_tests/python_only_compile.sh
|
||||
|
||||
- label: Basic Correctness Test # 30min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
fast_check: true
|
||||
torch_nightly: true
|
||||
source_file_dependencies:
|
||||
@ -88,8 +98,17 @@ steps:
|
||||
- pytest -v -s basic_correctness/test_cpu_offload.py
|
||||
- VLLM_TEST_ENABLE_ARTIFICIAL_PREEMPT=1 pytest -v -s basic_correctness/test_preemption.py
|
||||
|
||||
- label: Chunked Prefill Test
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/basic_correctness/test_chunked_prefill
|
||||
commands:
|
||||
- VLLM_ATTENTION_BACKEND=XFORMERS pytest -v -s basic_correctness/test_chunked_prefill.py
|
||||
- VLLM_ATTENTION_BACKEND=FLASH_ATTN pytest -v -s basic_correctness/test_chunked_prefill.py
|
||||
|
||||
- label: Core Test # 10min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
fast_check: true
|
||||
source_file_dependencies:
|
||||
- vllm/core
|
||||
@ -98,7 +117,7 @@ steps:
|
||||
commands:
|
||||
- pytest -v -s core
|
||||
|
||||
- label: Entrypoints Test (LLM) # 40min
|
||||
- label: Entrypoints Test # 40min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
fast_check: true
|
||||
@ -106,28 +125,19 @@ steps:
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/entrypoints/llm
|
||||
- tests/entrypoints/openai
|
||||
- tests/entrypoints/test_chat_utils
|
||||
- tests/entrypoints/offline_mode
|
||||
commands:
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_lazy_outlines.py --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py
|
||||
- pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_lazy_outlines.py --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_generate_multiple_loras.py --ignore=entrypoints/llm/test_guided_generate.py --ignore=entrypoints/llm/test_collective_rpc.py
|
||||
- pytest -v -s entrypoints/llm/test_lazy_outlines.py # it needs a clean process
|
||||
- pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process
|
||||
- VLLM_USE_V1=0 pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
|
||||
|
||||
- label: Entrypoints Test (API Server) # 40min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
fast_check: true
|
||||
torch_nightly: true
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/entrypoints/openai
|
||||
- tests/entrypoints/test_chat_utils
|
||||
commands:
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/openai/test_collective_rpc.py # PYTHONPATH is needed to import custom Worker extension
|
||||
- pytest -v -s entrypoints/openai --ignore=entrypoints/openai/test_chat_with_tool_reasoning.py --ignore=entrypoints/openai/test_oot_registration.py --ignore=entrypoints/openai/test_tensorizer_entrypoint.py --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/test_collective_rpc.py
|
||||
- pytest -v -s entrypoints/llm/test_generate_multiple_loras.py # it needs a clean process
|
||||
- VLLM_USE_V1=0 pytest -v -s entrypoints/llm/test_guided_generate.py # it needs a clean process
|
||||
- pytest -v -s entrypoints/openai --ignore=entrypoints/openai/test_chat_with_tool_reasoning.py --ignore=entrypoints/openai/test_oot_registration.py --ignore=entrypoints/openai/test_tensorizer_entrypoint.py --ignore=entrypoints/openai/correctness/
|
||||
- pytest -v -s entrypoints/test_chat_utils.py
|
||||
- VLLM_USE_V1=0 pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
|
||||
|
||||
- label: Distributed Tests (4 GPUs) # 10min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
@ -139,14 +149,13 @@ steps:
|
||||
- tests/distributed/test_utils
|
||||
- tests/distributed/test_pynccl
|
||||
- tests/distributed/test_events
|
||||
- tests/spec_decode/e2e/test_integration_dist_tp4
|
||||
- tests/compile/test_basic_correctness
|
||||
- examples/offline_inference/rlhf.py
|
||||
- examples/offline_inference/rlhf_colocate.py
|
||||
- tests/examples/offline_inference/data_parallel.py
|
||||
- tests/v1/test_async_llm_dp.py
|
||||
- tests/v1/test_external_lb_dp.py
|
||||
- tests/v1/test_internal_lb_dp.py
|
||||
- tests/v1/test_hybrid_lb_dp.py
|
||||
- tests/v1/engine/test_engine_core_client.py
|
||||
commands:
|
||||
# test with tp=2 and external_dp=2
|
||||
@ -158,13 +167,12 @@ steps:
|
||||
- python3 ../examples/offline_inference/data_parallel.py --enforce-eager
|
||||
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/test_async_llm_dp.py
|
||||
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/test_external_lb_dp.py
|
||||
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/test_internal_lb_dp.py
|
||||
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/test_hybrid_lb_dp.py
|
||||
- pytest -v -s v1/engine/test_engine_core_client.py::test_kv_cache_events_dp
|
||||
- pytest -v -s distributed/test_utils.py
|
||||
- pytest -v -s compile/test_basic_correctness.py
|
||||
- pytest -v -s distributed/test_pynccl.py
|
||||
- pytest -v -s distributed/test_events.py
|
||||
- pytest -v -s spec_decode/e2e/test_integration_dist_tp4.py
|
||||
# TODO: create a dedicated test section for multi-GPU example tests
|
||||
# when we have multiple distributed example tests
|
||||
- pushd ../examples/offline_inference
|
||||
@ -190,7 +198,7 @@ steps:
|
||||
- pytest -v -s distributed/test_eplb_execute.py
|
||||
|
||||
- label: Metrics, Tracing Test # 10min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
num_gpus: 2
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
@ -233,26 +241,7 @@ steps:
|
||||
# OOM in the CI unless we run this separately
|
||||
- pytest -v -s tokenization
|
||||
|
||||
- label: V1 Test e2e + engine
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/v1
|
||||
commands:
|
||||
# TODO: accuracy does not match, whether setting
|
||||
# VLLM_USE_FLASHINFER_SAMPLER or not on H100.
|
||||
- pytest -v -s v1/e2e
|
||||
- pytest -v -s v1/engine
|
||||
|
||||
- label: V1 Test entrypoints
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/v1
|
||||
commands:
|
||||
- pytest -v -s v1/entrypoints
|
||||
|
||||
- label: V1 Test others
|
||||
- label: V1 Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
@ -260,20 +249,22 @@ steps:
|
||||
commands:
|
||||
# split the test to avoid interference
|
||||
- pytest -v -s v1/core
|
||||
- pytest -v -s v1/executor
|
||||
- pytest -v -s v1/engine
|
||||
- pytest -v -s v1/entrypoints
|
||||
- pytest -v -s v1/sample
|
||||
- pytest -v -s v1/logits_processors
|
||||
- pytest -v -s v1/worker
|
||||
- pytest -v -s v1/structured_output
|
||||
- pytest -v -s v1/spec_decode
|
||||
- pytest -v -s v1/kv_connector/unit
|
||||
- pytest -v -s v1/metrics
|
||||
- pytest -v -s v1/test_serial_utils.py
|
||||
- pytest -v -s v1/test_utils.py
|
||||
- pytest -v -s v1/test_oracle.py
|
||||
- pytest -v -s v1/test_metrics_reader.py
|
||||
# TODO: accuracy does not match, whether setting
|
||||
# VLLM_USE_FLASHINFER_SAMPLER or not on H100.
|
||||
- pytest -v -s v1/e2e
|
||||
# Integration test for streaming correctness (requires special branch).
|
||||
- pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api
|
||||
- pip install -U git+https://github.com/robertgshaw2-neuralmagic/lm-evaluation-harness.git@streaming-api
|
||||
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
|
||||
|
||||
- label: Examples Test # 25min
|
||||
@ -291,7 +282,7 @@ steps:
|
||||
- python3 offline_inference/llm_engine_example.py
|
||||
- python3 offline_inference/audio_language.py --seed 0
|
||||
- python3 offline_inference/vision_language.py --seed 0
|
||||
- python3 offline_inference/vision_language_pooling.py --seed 0
|
||||
- python3 offline_inference/vision_language_embedding.py --seed 0
|
||||
- python3 offline_inference/vision_language_multi_image.py --seed 0
|
||||
- VLLM_USE_V1=0 python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors
|
||||
- python3 offline_inference/encoder_decoder.py
|
||||
@ -301,6 +292,15 @@ steps:
|
||||
- python3 offline_inference/basic/score.py
|
||||
- VLLM_USE_V1=0 python3 offline_inference/profiling.py --model facebook/opt-125m run_num_steps --num-steps 2
|
||||
|
||||
- label: Prefix Caching Test # 9min
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/prefix_caching
|
||||
commands:
|
||||
- pytest -v -s prefix_caching
|
||||
|
||||
|
||||
- label: Platform Tests (CUDA)
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
@ -320,12 +320,23 @@ steps:
|
||||
- pytest -v -s samplers
|
||||
- VLLM_USE_FLASHINFER_SAMPLER=1 pytest -v -s samplers
|
||||
|
||||
- label: LoRA Test %N # 15min each
|
||||
- label: Speculative decoding tests # 40min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
- vllm/spec_decode
|
||||
- tests/spec_decode
|
||||
- vllm/model_executor/models/eagle.py
|
||||
commands:
|
||||
- pytest -v -s spec_decode/e2e/test_multistep_correctness.py
|
||||
- VLLM_ATTENTION_BACKEND=FLASH_ATTN pytest -v -s spec_decode --ignore=spec_decode/e2e/test_multistep_correctness.py --ignore=spec_decode/e2e/test_mtp_correctness.py
|
||||
- pytest -v -s spec_decode/e2e/test_eagle_correctness.py
|
||||
|
||||
- label: LoRA Test %N # 15min each
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- vllm/lora
|
||||
- tests/lora
|
||||
command: pytest -v -s lora --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --ignore=lora/test_chatglm3_tp.py --ignore=lora/test_llama_tp.py --ignore=lora/test_llm_with_multi_loras.py
|
||||
command: pytest -v -s lora --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --ignore=lora/test_chatglm3_tp.py --ignore=lora/test_llama_tp.py
|
||||
parallelism: 4
|
||||
|
||||
- label: PyTorch Compilation Unit Tests
|
||||
@ -341,11 +352,9 @@ steps:
|
||||
- pytest -v -s compile/test_silu_mul_quant_fusion.py
|
||||
- pytest -v -s compile/test_sequence_parallelism.py
|
||||
- pytest -v -s compile/test_async_tp.py
|
||||
- pytest -v -s compile/test_fusion_all_reduce.py
|
||||
- pytest -v -s compile/test_decorator.py
|
||||
|
||||
- label: PyTorch Fullgraph Smoke Test # 9min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
torch_nightly: true
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
@ -356,10 +365,9 @@ steps:
|
||||
- pytest -v -s compile/piecewise/test_simple.py
|
||||
- pytest -v -s compile/piecewise/test_toy_llama.py
|
||||
- pytest -v -s compile/piecewise/test_full_cudagraph.py
|
||||
- pytest -v -s compile/piecewise/test_multiple_graphs.py
|
||||
|
||||
- label: PyTorch Fullgraph Test # 18min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
torch_nightly: true
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
@ -368,7 +376,7 @@ steps:
|
||||
- pytest -v -s compile/test_full_graph.py
|
||||
|
||||
- label: Kernels Core Operation Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- tests/kernels/core
|
||||
@ -376,7 +384,7 @@ steps:
|
||||
- pytest -v -s kernels/core
|
||||
|
||||
- label: Kernels Attention Test %N
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- csrc/attention/
|
||||
- vllm/attention
|
||||
@ -387,26 +395,23 @@ steps:
|
||||
parallelism: 2
|
||||
|
||||
- label: Kernels Quantization Test %N
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- csrc/quantization/
|
||||
- vllm/model_executor/layers/quantization
|
||||
- tests/kernels/quantization
|
||||
commands:
|
||||
- pytest -v -s kernels/quantization --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
- pytest -v -s kernels/quantization --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
parallelism: 2
|
||||
|
||||
- label: Kernels MoE Test %N
|
||||
- label: Kernels MoE Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
source_file_dependencies:
|
||||
- csrc/quantization/cutlass_w8a8/moe/
|
||||
- csrc/moe/
|
||||
- tests/kernels/moe
|
||||
- vllm/model_executor/layers/fused_moe/
|
||||
- vllm/distributed/device_communicators/
|
||||
commands:
|
||||
- pytest -v -s kernels/moe --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
parallelism: 2
|
||||
- pytest -v -s kernels/moe
|
||||
|
||||
- label: Kernels Mamba Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
@ -418,6 +423,7 @@ steps:
|
||||
|
||||
- label: Tensorizer Test # 11min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
soft_fail: true
|
||||
source_file_dependencies:
|
||||
- vllm/model_executor/model_loader
|
||||
- tests/tensorizer_loader
|
||||
@ -429,7 +435,8 @@ steps:
|
||||
- pytest -v -s entrypoints/openai/test_tensorizer_entrypoint.py
|
||||
|
||||
- label: Model Executor Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
soft_fail: true
|
||||
source_file_dependencies:
|
||||
- vllm/model_executor
|
||||
- tests/model_executor
|
||||
@ -439,7 +446,7 @@ steps:
|
||||
- pytest -v -s model_executor
|
||||
|
||||
- label: Benchmarks # 9min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
working_dir: "/vllm-workspace/.buildkite"
|
||||
source_file_dependencies:
|
||||
- benchmarks/
|
||||
@ -447,7 +454,7 @@ steps:
|
||||
- bash scripts/run-benchmarks.sh
|
||||
|
||||
- label: Benchmarks CLI Test # 10min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/benchmarks/
|
||||
@ -462,17 +469,19 @@ steps:
|
||||
- tests/quantization
|
||||
commands:
|
||||
# temporary install here since we need nightly, will move to requirements/test.in
|
||||
# after torchao 0.12 release, and pin a working version of torchao nightly here
|
||||
- pip install --pre torchao==0.13.0.dev20250814 --index-url https://download.pytorch.org/whl/nightly/cu128
|
||||
# after torchao 0.12 release
|
||||
- pip install --pre torchao --index-url https://download.pytorch.org/whl/nightly/cu126
|
||||
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization
|
||||
|
||||
- label: LM Eval Small Models # 53min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
working_dir: "/vllm-workspace/.buildkite/lm-eval-harness"
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- vllm/model_executor/layers/quantization
|
||||
commands:
|
||||
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-small.txt --tp-size=1
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-small.txt --tp-size=1
|
||||
|
||||
- label: OpenAI API correctness
|
||||
mirror_hardwares: [amdexperimental]
|
||||
@ -524,6 +533,8 @@ steps:
|
||||
- vllm/
|
||||
- tests/models/language
|
||||
commands:
|
||||
# Install causal-conv1d for plamo2 models here, as it is not compatible with pip-compile.
|
||||
- pip install 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.0.post8'
|
||||
- pip freeze | grep -E 'torch'
|
||||
- pytest -v -s models/language -m core_model
|
||||
|
||||
@ -534,10 +545,8 @@ steps:
|
||||
- vllm/
|
||||
- tests/models/language/generation
|
||||
commands:
|
||||
# Install fast path packages for testing against transformers
|
||||
# Note: also needed to run plamo2 model in vLLM
|
||||
- uv pip install --system --no-build-isolation 'git+https://github.com/state-spaces/mamba@v2.2.5'
|
||||
- uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2'
|
||||
# Install causal-conv1d for plamo2 models here, as it is not compatible with pip-compile.
|
||||
- pip install 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.0.post8'
|
||||
- pytest -v -s models/language/generation -m hybrid_model
|
||||
|
||||
- label: Language Models Test (Extended Generation) # 1hr20min
|
||||
@ -560,14 +569,6 @@ steps:
|
||||
commands:
|
||||
- pytest -v -s models/language/pooling -m 'not core_model'
|
||||
|
||||
- label: Multi-Modal Processor Test
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/models/multimodal
|
||||
commands:
|
||||
- pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
|
||||
- pytest -v -s models/multimodal/processing
|
||||
|
||||
- label: Multi-Modal Models Test (Standard)
|
||||
mirror_hardwares: [amdexperimental]
|
||||
torch_nightly: true
|
||||
@ -577,7 +578,8 @@ steps:
|
||||
commands:
|
||||
- pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
|
||||
- pip freeze | grep -E 'torch'
|
||||
- pytest -v -s models/multimodal -m core_model --ignore models/multimodal/generation/test_whisper.py --ignore models/multimodal/processing
|
||||
- pytest -v -s models/multimodal/processing
|
||||
- pytest -v -s --ignore models/multimodal/generation/test_whisper.py models/multimodal -m core_model
|
||||
- cd .. && pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model # Otherwise, mp_method="spawn" doesn't work
|
||||
|
||||
- label: Multi-Modal Models Test (Extended) 1
|
||||
@ -588,7 +590,7 @@ steps:
|
||||
- tests/models/multimodal
|
||||
commands:
|
||||
- pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
|
||||
- pytest -v -s models/multimodal -m 'not core_model' --ignore models/multimodal/generation/test_common.py --ignore models/multimodal/processing
|
||||
- pytest -v -s --ignore models/multimodal/generation/test_common.py --ignore models/multimodal/processing models/multimodal -m 'not core_model'
|
||||
|
||||
- label: Multi-Modal Models Test (Extended) 2
|
||||
mirror_hardwares: [amdexperimental]
|
||||
@ -611,7 +613,7 @@ steps:
|
||||
- pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model'
|
||||
|
||||
- label: Quantized Models Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
source_file_dependencies:
|
||||
- vllm/model_executor/layers/quantization
|
||||
- tests/models/quantization
|
||||
@ -620,7 +622,7 @@ steps:
|
||||
|
||||
# This test is used only in PR development phase to test individual models and should never run on main
|
||||
- label: Custom Models Test
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
optional: true
|
||||
commands:
|
||||
- echo 'Testing custom models...'
|
||||
@ -628,61 +630,11 @@ steps:
|
||||
# e.g. pytest -v -s models/encoder_decoder/vision_language/test_mllama.py
|
||||
# *To avoid merge conflicts, remember to REMOVE (not just comment out) them before merging the PR*
|
||||
|
||||
- label: Transformers Nightly Models Test
|
||||
working_dir: "/vllm-workspace/"
|
||||
optional: true
|
||||
commands:
|
||||
- pip install --upgrade git+https://github.com/huggingface/transformers
|
||||
- pytest -v -s tests/models/test_initialization.py
|
||||
- pytest -v -s tests/models/multimodal/processing/
|
||||
- pytest -v -s tests/models/multimodal/test_mapping.py
|
||||
- python3 examples/offline_inference/basic/chat.py
|
||||
- python3 examples/offline_inference/audio_language.py --model-type whisper
|
||||
- python3 examples/offline_inference/vision_language.py --model-type qwen2_5_vl
|
||||
|
||||
- label: Blackwell Test
|
||||
working_dir: "/vllm-workspace/"
|
||||
gpu: b200
|
||||
# optional: true
|
||||
source_file_dependencies:
|
||||
- csrc/quantization/fp4/
|
||||
- csrc/attention/mla/
|
||||
- csrc/quantization/cutlass_w8a8/moe/
|
||||
- vllm/model_executor/layers/fused_moe/cutlass_moe.py
|
||||
- vllm/model_executor/layers/fused_moe/flashinfer_cutlass_moe.py
|
||||
- vllm/model_executor/layers/fused_moe/flashinfer_cutlass_prepare_finalize.py
|
||||
- vllm/model_executor/layers/quantization/utils/flashinfer_utils.py
|
||||
- vllm/v1/attention/backends/flashinfer.py
|
||||
- vllm/compilation/fusion.py
|
||||
- vllm/compilation/fusion_attn.py
|
||||
commands:
|
||||
- nvidia-smi
|
||||
- python3 examples/offline_inference/basic/chat.py
|
||||
# Attention
|
||||
# num_heads2 broken by https://github.com/flashinfer-ai/flashinfer/issues/1353
|
||||
- pytest -v -s tests/kernels/attention/test_flashinfer.py -k 'not num_heads2'
|
||||
- pytest -v -s tests/kernels/attention/test_flashinfer_trtllm_attention.py
|
||||
- pytest -v -s tests/kernels/test_cutlass_mla_decode.py
|
||||
# Quantization
|
||||
- pytest -v -s tests/kernels/quantization/test_cutlass_scaled_mm.py -k 'fp8'
|
||||
- pytest -v -s tests/kernels/quantization/test_nvfp4_quant.py
|
||||
# - pytest -v -s tests/kernels/quantization/test_silu_nvfp4_quant_fusion.py
|
||||
- pytest -v -s tests/kernels/quantization/test_nvfp4_scaled_mm.py
|
||||
- pytest -v -s tests/kernels/quantization/test_flashinfer_scaled_mm.py
|
||||
- pytest -v -s tests/kernels/quantization/test_flashinfer_nvfp4_scaled_mm.py
|
||||
- pytest -v -s tests/kernels/moe/test_nvfp4_moe.py
|
||||
- pytest -v -s tests/kernels/moe/test_mxfp4_moe.py
|
||||
# Fusion
|
||||
- pytest -v -s tests/compile/test_fusion_all_reduce.py
|
||||
- pytest -v -s tests/compile/test_fusion_attn.py::test_attention_quant_pattern
|
||||
- pytest -v -s tests/kernels/moe/test_flashinfer.py
|
||||
# - pytest -v -s tests/compile/test_silu_mul_quant_fusion.py
|
||||
|
||||
##### 1 GPU test #####
|
||||
##### multi gpus test #####
|
||||
|
||||
- label: Distributed Comm Ops Test # 7min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
num_gpus: 2
|
||||
source_file_dependencies:
|
||||
@ -752,9 +704,10 @@ steps:
|
||||
- pytest -v -s distributed/test_sequence_parallel.py
|
||||
# this test fails consistently.
|
||||
# TODO: investigate and fix
|
||||
# - pytest -v -s spec_decode/e2e/test_integration_dist_tp2.py
|
||||
- VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py
|
||||
- VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s kv_transfer/test_disagg.py
|
||||
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown
|
||||
- pytest -v -s models/multimodal/generation/test_maverick.py
|
||||
|
||||
- label: Plugin Tests (2 GPUs) # 40min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
@ -769,11 +722,6 @@ steps:
|
||||
- pytest -v -s plugins_tests/test_platform_plugins.py
|
||||
- pip uninstall vllm_add_dummy_platform -y
|
||||
# end platform plugin tests
|
||||
# begin io_processor plugins test, all the code in between uses the prithvi_io_processor plugin
|
||||
- pip install -e ./plugins/prithvi_io_processor_plugin
|
||||
- pytest -v -s plugins_tests/test_io_processor_plugins.py
|
||||
- pip uninstall prithvi_io_processor_plugin -y
|
||||
# end io_processor plugins test
|
||||
# other tests continue here:
|
||||
- pytest -v -s plugins_tests/test_scheduler_plugins.py
|
||||
- pip install -e ./plugins/vllm_add_dummy_model
|
||||
@ -782,8 +730,29 @@ steps:
|
||||
- pytest -v -s models/test_oot_registration.py # it needs a clean process
|
||||
- pytest -v -s plugins/lora_resolvers # unit tests for in-tree lora resolver plugins
|
||||
|
||||
- label: Multi-step Tests (4 GPUs) # 36min
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
num_gpus: 4
|
||||
source_file_dependencies:
|
||||
- vllm/model_executor/layers/sampler.py
|
||||
- vllm/sequence.py
|
||||
- vllm/worker/worker_base.py
|
||||
- vllm/worker/worker.py
|
||||
- vllm/worker/multi_step_worker.py
|
||||
- vllm/worker/model_runner_base.py
|
||||
- vllm/worker/model_runner.py
|
||||
- vllm/worker/multi_step_model_runner.py
|
||||
- vllm/engine
|
||||
- tests/multi_step
|
||||
commands:
|
||||
# this test is quite flaky
|
||||
# TODO: investigate and fix.
|
||||
# - pytest -v -s multi_step/test_correctness_async_llm.py
|
||||
- pytest -v -s multi_step/test_correctness_llm.py
|
||||
|
||||
- label: Pipeline Parallelism Test # 45min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
num_gpus: 4
|
||||
source_file_dependencies:
|
||||
@ -797,7 +766,7 @@ steps:
|
||||
- pytest -v -s distributed/test_pipeline_parallel.py
|
||||
|
||||
- label: LoRA TP Test (Distributed)
|
||||
mirror_hardwares: [amdexperimental]
|
||||
mirror_hardwares: [amdexperimental, amdproduction]
|
||||
num_gpus: 4
|
||||
source_file_dependencies:
|
||||
- vllm/lora
|
||||
@ -810,14 +779,12 @@ steps:
|
||||
# requires multi-GPU testing for validation.
|
||||
- pytest -v -s -x lora/test_chatglm3_tp.py
|
||||
- pytest -v -s -x lora/test_llama_tp.py
|
||||
- pytest -v -s -x lora/test_llm_with_multi_loras.py
|
||||
|
||||
|
||||
- label: Weight Loading Multiple GPU Test # 33min
|
||||
mirror_hardwares: [amdexperimental]
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
num_gpus: 2
|
||||
optional: true
|
||||
num_gpus: 2
|
||||
source_file_dependencies:
|
||||
- vllm/
|
||||
- tests/weight_loading
|
||||
@ -865,10 +832,3 @@ steps:
|
||||
commands:
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large.txt --tp-size=4
|
||||
|
||||
- label: Qwen MoE EP Test # optional
|
||||
gpu: h200
|
||||
optional: true
|
||||
num_gpus: 2
|
||||
commands:
|
||||
- CUDA_VISIBLE_DEVICES=1,2 VLLM_ALL2ALL_BACKEND=deepep_high_throughput VLLM_USE_DEEP_GEMM=1 VLLM_LOGGING_LEVEL=DEBUG python3 /vllm-workspace/examples/offline_inference/data_parallel.py --model Qwen/Qwen1.5-MoE-A2.7B --tp-size=1 --dp-size=2 --max-model-len 2048
|
||||
|
||||
@ -1,6 +0,0 @@
|
||||
# https://developers.google.com/gemini-code-assist/docs/customize-gemini-behavior-github
|
||||
have_fun: false # Just review the code
|
||||
code_review:
|
||||
comment_severity_threshold: HIGH # Reduce quantity of comments
|
||||
pull_request_opened:
|
||||
summary: false # Don't summarize the PR in a separate comment
|
||||
58
.github/CODEOWNERS
vendored
58
.github/CODEOWNERS
vendored
@ -9,80 +9,46 @@
|
||||
/vllm/worker/worker_base.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/worker/worker.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/model_executor/layers/sampler.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/model_executor/layers/quantization @mgoin @robertgshaw2-redhat @tlrmchlsmth @yewentao256
|
||||
/vllm/model_executor/layers/mamba @tdoublep
|
||||
/vllm/model_executor/layers/quantization @mgoin @robertgshaw2-redhat @tlrmchlsmth
|
||||
/vllm/model_executor/guided_decoding @mgoin @russellb @aarnphm
|
||||
/vllm/multimodal @DarkLight1337 @ywang96
|
||||
/vllm/vllm_flash_attn @LucasWilkinson
|
||||
/vllm/lora @jeejeelee
|
||||
/vllm/reasoning @aarnphm
|
||||
/vllm/entrypoints @aarnphm
|
||||
/vllm/compilation @zou3519 @youkaichao @ProExpertProg
|
||||
CMakeLists.txt @tlrmchlsmth @LucasWilkinson
|
||||
CMakeLists.txt @tlrmchlsmth
|
||||
|
||||
# Any change to the VllmConfig changes can have a large user-facing impact,
|
||||
# so spam a lot of people
|
||||
/vllm/config @simon-mo @WoosukKwon @youkaichao @robertgshaw2-redhat @mgoin @tlrmchlsmth @houseroad @hmellor @yewentao256 @ProExpertProg
|
||||
/vllm/config.py @simon-mo @WoosukKwon @youkaichao @robertgshaw2-redhat @mgoin @tlrmchlsmth @houseroad @hmellor
|
||||
|
||||
# vLLM V1
|
||||
/vllm/v1 @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat
|
||||
/vllm/v1/structured_output @mgoin @russellb @aarnphm
|
||||
/vllm/v1/attention/backends/triton_attn.py @tdoublep
|
||||
|
||||
# Test ownership
|
||||
/.buildkite/lm-eval-harness @mgoin @simon-mo
|
||||
/tests/async_engine @njhill @robertgshaw2-redhat @simon-mo
|
||||
/tests/basic_correctness/test_chunked_prefill @rkooo567 @comaniac
|
||||
/tests/distributed/test_multi_node_assignment.py @youkaichao
|
||||
/tests/distributed/test_pipeline_parallel.py @youkaichao
|
||||
/tests/distributed/test_same_node.py @youkaichao
|
||||
/tests/entrypoints @DarkLight1337 @robertgshaw2-redhat @simon-mo @aarnphm
|
||||
/tests/kernels @tlrmchlsmth @WoosukKwon @yewentao256
|
||||
/tests/entrypoints/llm/test_guided_generate.py @mgoin @russellb @aarnphm
|
||||
/tests/kernels @tlrmchlsmth @WoosukKwon
|
||||
/tests/model_executor/test_guided_processors.py @mgoin @russellb
|
||||
/tests/models @DarkLight1337 @ywang96
|
||||
/tests/multi_step @alexm-redhat @comaniac
|
||||
/tests/multimodal @DarkLight1337 @ywang96
|
||||
/tests/prefix_caching @comaniac @KuntaiDu
|
||||
/tests/quantization @mgoin @robertgshaw2-redhat @yewentao256
|
||||
/tests/quantization @mgoin @robertgshaw2-redhat
|
||||
/tests/spec_decode @njhill @LiuXiaoxuanPKU
|
||||
/tests/test_inputs.py @DarkLight1337 @ywang96
|
||||
/tests/v1/entrypoints/llm/test_struct_output_generate.py @mgoin @russellb @aarnphm
|
||||
/tests/v1/structured_output @mgoin @russellb @aarnphm
|
||||
/tests/weight_loading @mgoin @youkaichao @yewentao256
|
||||
/tests/weight_loading @mgoin @youkaichao
|
||||
/tests/lora @jeejeelee
|
||||
/tests/models/language/generation/test_hybrid.py @tdoublep
|
||||
|
||||
# Docs
|
||||
/docs @hmellor
|
||||
mkdocs.yaml @hmellor
|
||||
|
||||
# CPU
|
||||
/vllm/v1/worker/^cpu @bigPYJ1151
|
||||
/csrc/cpu @bigPYJ1151
|
||||
/vllm/platforms/cpu.py @bigPYJ1151
|
||||
/cmake/cpu_extension.cmake @bigPYJ1151
|
||||
/docker/Dockerfile.cpu @bigPYJ1151
|
||||
|
||||
# Intel GPU
|
||||
/vllm/v1/worker/^xpu @jikunshang
|
||||
/vllm/platforms/xpu.py @jikunshang
|
||||
/docker/Dockerfile.xpu @jikunshang
|
||||
|
||||
# Qwen-specific files
|
||||
/vllm/attention/backends/dual_chunk_flash_attn.py @sighingnow
|
||||
/vllm/model_executor/models/qwen* @sighingnow
|
||||
|
||||
# Mistral-specific files
|
||||
/vllm/model_executor/models/mistral*.py @patrickvonplaten
|
||||
/vllm/model_executor/models/mixtral*.py @patrickvonplaten
|
||||
/vllm/model_executor/models/voxtral*.py @patrickvonplaten
|
||||
/vllm/model_executor/models/pixtral*.py @patrickvonplaten
|
||||
/vllm/transformers_utils/configs/mistral.py @patrickvonplaten
|
||||
/vllm/transformers_utils/tokenizers/mistral.py @patrickvonplaten
|
||||
|
||||
# Kernels
|
||||
/vllm/attention/ops/chunked_prefill_paged_decode.py @tdoublep
|
||||
/vllm/attention/ops/triton_unified_attention.py @tdoublep
|
||||
|
||||
# ROCm related: specify owner with write access to notify AMD folks for careful code review
|
||||
/docker/Dockerfile.rocm* @gshtras
|
||||
/vllm/v1/attention/backends/rocm*.py @gshtras
|
||||
/vllm/v1/attention/backends/mla/rocm*.py @gshtras
|
||||
/vllm/attention/ops/rocm*.py @gshtras
|
||||
/vllm/model_executor/layers/fused_moe/rocm*.py @gshtras
|
||||
|
||||
|
||||
2
.github/ISSUE_TEMPLATE/750-RFC.yml
vendored
2
.github/ISSUE_TEMPLATE/750-RFC.yml
vendored
@ -46,7 +46,7 @@ body:
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: >
|
||||
Thanks for contributing 🎉! The vLLM core team hosts a biweekly RFC review session at 9:30AM Pacific Time, while most RFCs can be discussed online, you can optionally sign up for a slot to discuss your RFC online [here](https://docs.google.com/document/d/1CiLVBZeIVfR7_PNAKVSusxpceywkoOOB78qoWqHvSZc/edit).
|
||||
Thanks for contributing 🎉!
|
||||
- type: checkboxes
|
||||
id: askllm
|
||||
attributes:
|
||||
|
||||
21
.github/PULL_REQUEST_TEMPLATE.md
vendored
21
.github/PULL_REQUEST_TEMPLATE.md
vendored
@ -1,5 +1,10 @@
|
||||
<!-- markdownlint-disable -->
|
||||
PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED.
|
||||
## Essential Elements of an Effective PR Description Checklist
|
||||
- [ ] The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
|
||||
- [ ] The test plan, such as providing test command.
|
||||
- [ ] The test results, such as pasting the results comparison before and after, or e2e results
|
||||
- [ ] (Optional) The necessary documentation update, such as updating `supported_models.md` and `examples` for a new model.
|
||||
|
||||
PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS ABOVE HAVE BEEN CONSIDERED.
|
||||
|
||||
## Purpose
|
||||
|
||||
@ -7,15 +12,7 @@ PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTT
|
||||
|
||||
## Test Result
|
||||
|
||||
---
|
||||
<details>
|
||||
<summary> Essential Elements of an Effective PR Description Checklist </summary>
|
||||
|
||||
- [ ] The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
|
||||
- [ ] The test plan, such as providing test command.
|
||||
- [ ] The test results, such as pasting the results comparison before and after, or e2e results
|
||||
- [ ] (Optional) The necessary documentation update, such as updating `supported_models.md` and `examples` for a new model.
|
||||
- [ ] (Optional) Release notes update. If your change is user facing, please update the release notes draft in the [Google Doc](https://docs.google.com/document/d/1YyVqrgX4gHTtrstbq8oWUImOyPCKSGnJ7xtTpmXzlRs/edit?tab=t.0).
|
||||
</details>
|
||||
## (Optional) Documentation Update
|
||||
|
||||
<!--- pyml disable-next-line no-emphasis-as-heading -->
|
||||
**BEFORE SUBMITTING, PLEASE READ <https://docs.vllm.ai/en/latest/contributing>** (anything written below this line will be removed by GitHub Actions)
|
||||
|
||||
22
.github/mergify.yml
vendored
22
.github/mergify.yml
vendored
@ -86,6 +86,8 @@ pull_request_rules:
|
||||
- and:
|
||||
- files~=^vllm/model_executor/models/
|
||||
- files=vllm/model_executor/models/registry.py
|
||||
- files=tests/models/registry.py
|
||||
- files=docs/models/supported_models.md
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
@ -118,20 +120,6 @@ pull_request_rules:
|
||||
add:
|
||||
- qwen
|
||||
|
||||
- name: label-gpt-oss
|
||||
description: Automatically apply gpt-oss label
|
||||
conditions:
|
||||
- or:
|
||||
- files~=^examples/.*gpt[-_]?oss.*\.py
|
||||
- files~=^tests/.*gpt[-_]?oss.*\.py
|
||||
- files~=^vllm/model_executor/models/.*gpt[-_]?oss.*\.py
|
||||
- files~=^vllm/model_executor/layers/.*gpt[-_]?oss.*\.py
|
||||
- title~=(?i)gpt[-_]?oss
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- gpt-oss
|
||||
|
||||
- name: label-rocm
|
||||
description: Automatically apply rocm label
|
||||
conditions:
|
||||
@ -163,6 +151,9 @@ pull_request_rules:
|
||||
- files=examples/offline_inference/structured_outputs.py
|
||||
- files=examples/online_serving/openai_chat_completion_structured_outputs.py
|
||||
- files=examples/online_serving/openai_chat_completion_structured_outputs_with_reasoning.py
|
||||
- files~=^vllm/model_executor/guided_decoding/
|
||||
- files=tests/model_executor/test_guided_processors.py
|
||||
- files=tests/entrypoints/llm/test_guided_generate.py
|
||||
- files~=^tests/v1/structured_output/
|
||||
- files=tests/v1/entrypoints/llm/test_guided_generate.py
|
||||
- files~=^vllm/v1/structured_output/
|
||||
@ -175,7 +166,10 @@ pull_request_rules:
|
||||
description: Automatically apply speculative-decoding label
|
||||
conditions:
|
||||
- or:
|
||||
- files~=^vllm/spec_decode/
|
||||
- files~=^vllm/v1/spec_decode/
|
||||
- files=vllm/model_executor/layers/spec_decode_base_sampler.py
|
||||
- files~=^tests/spec_decode/
|
||||
- files~=^tests/v1/spec_decode/
|
||||
- files~=^examples/.*(spec_decode|mlpspeculator|eagle|speculation).*\.py
|
||||
- files~=^vllm/model_executor/models/.*eagle.*\.py
|
||||
|
||||
21
.github/scale-config.yml
vendored
21
.github/scale-config.yml
vendored
@ -1,21 +0,0 @@
|
||||
# scale-config.yml:
|
||||
# Powers what instance types are available for GHA auto-scaled
|
||||
# runners. Runners listed here will be available as self hosted
|
||||
# runners, configuration is directly pulled from the main branch.
|
||||
# runner_types:
|
||||
# runner_label:
|
||||
# instance_type: m4.large
|
||||
# os: linux
|
||||
# # min_available defaults to the global cfg in the ALI Terraform
|
||||
# min_available: undefined
|
||||
# # when max_available value is not defined, no max runners is enforced
|
||||
# max_available: undefined
|
||||
# disk_size: 50
|
||||
# is_ephemeral: true
|
||||
|
||||
runner_types:
|
||||
linux.2xlarge:
|
||||
disk_size: 150
|
||||
instance_type: c5.2xlarge
|
||||
is_ephemeral: true
|
||||
os: linux
|
||||
8
.github/scripts/cleanup_pr_body.sh
vendored
8
.github/scripts/cleanup_pr_body.sh
vendored
@ -15,11 +15,11 @@ NEW=/tmp/new_pr_body.txt
|
||||
gh pr view --json body --template "{{.body}}" "${PR_NUMBER}" > "${OLD}"
|
||||
cp "${OLD}" "${NEW}"
|
||||
|
||||
# Remove markdown comments (like the <!-- markdownlint-disable --> at the start)
|
||||
sed -i '/<!--.*-->$/d' "${NEW}"
|
||||
# Remove "FIX #xxxx (*link existing issues this PR will resolve*)"
|
||||
sed -i '/FIX #xxxx.*$/d' "${NEW}"
|
||||
|
||||
# Remove "PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED."
|
||||
sed -i '/PLEASE FILL IN THE PR DESCRIPTION HERE.*$/d' "${NEW}"
|
||||
# Remove "FILL IN THE PR DESCRIPTION HERE"
|
||||
sed -i '/FILL IN THE PR DESCRIPTION HERE/d' "${NEW}"
|
||||
|
||||
# Remove all lines after and including "**BEFORE SUBMITTING, PLEASE READ THE CHECKLIST BELOW AND FILL IN THE DESCRIPTION ABOVE**"
|
||||
sed -i '/\*\*BEFORE SUBMITTING, PLEASE READ.*\*\*/,$d' "${NEW}"
|
||||
|
||||
309
.github/workflows/issue_autolabel.yml
vendored
309
.github/workflows/issue_autolabel.yml
vendored
@ -1,309 +0,0 @@
|
||||
name: Label issues based on keywords
|
||||
on:
|
||||
issues:
|
||||
types: [opened, edited, reopened]
|
||||
permissions:
|
||||
issues: write # needed so the workflow can add labels
|
||||
contents: read
|
||||
concurrency:
|
||||
group: issue-labeler-${{ github.event.issue.number }}
|
||||
cancel-in-progress: true
|
||||
jobs:
|
||||
add-labels:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Label issues based on keywords
|
||||
uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
|
||||
with:
|
||||
script: |
|
||||
// Configuration: Add new labels and keywords here
|
||||
const labelConfig = {
|
||||
rocm: {
|
||||
// Keyword search - matches whole words only (with word boundaries)
|
||||
keywords: [
|
||||
{
|
||||
term: "composable kernel",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "rccl",
|
||||
searchIn: "body" // only search in body
|
||||
},
|
||||
{
|
||||
term: "migraphx",
|
||||
searchIn: "title" // only search in title
|
||||
},
|
||||
{
|
||||
term: "hipgraph",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "ROCm System Management Interface",
|
||||
searchIn: "body"
|
||||
},
|
||||
],
|
||||
|
||||
// Substring search - matches anywhere in text (partial matches)
|
||||
substrings: [
|
||||
{
|
||||
term: "VLLM_ROCM_",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "aiter",
|
||||
searchIn: "title"
|
||||
},
|
||||
{
|
||||
term: "rocm",
|
||||
searchIn: "title"
|
||||
},
|
||||
{
|
||||
term: "amd",
|
||||
searchIn: "title"
|
||||
},
|
||||
{
|
||||
term: "hip-",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "gfx",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "cdna",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "rdna",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "torch_hip",
|
||||
searchIn: "body" // only in body
|
||||
},
|
||||
{
|
||||
term: "_hip",
|
||||
searchIn: "both"
|
||||
},
|
||||
{
|
||||
term: "hip_",
|
||||
searchIn: "both"
|
||||
},
|
||||
|
||||
// ROCm tools and libraries
|
||||
{
|
||||
term: "hipify",
|
||||
searchIn: "both"
|
||||
},
|
||||
],
|
||||
|
||||
// Regex patterns - for complex pattern matching
|
||||
regexPatterns: [
|
||||
{
|
||||
pattern: "\\bmi\\d{3}[a-z]*\\b",
|
||||
description: "AMD GPU names (mi + 3 digits + optional letters)",
|
||||
flags: "gi",
|
||||
searchIn: "both" // "title", "body", or "both"
|
||||
}
|
||||
],
|
||||
},
|
||||
};
|
||||
|
||||
// Helper function to create regex based on search type
|
||||
function createSearchRegex(term, type) {
|
||||
// Escape special regex characters in the term
|
||||
const escapedTerm = term.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
|
||||
|
||||
switch (type) {
|
||||
case 'keyword':
|
||||
// Word boundary search - matches whole words only
|
||||
return new RegExp(`\\b${escapedTerm}\\b`, "gi");
|
||||
case 'substring':
|
||||
// Substring search - matches anywhere in the text
|
||||
return new RegExp(escapedTerm, "gi");
|
||||
default:
|
||||
throw new Error(`Unknown search type: ${type}`);
|
||||
}
|
||||
}
|
||||
|
||||
// Helper function to find matching terms in text with line information
|
||||
function findMatchingTermsWithLines(text, searchTerms = [], searchType = 'keyword', searchLocation = '') {
|
||||
const matches = [];
|
||||
const lines = text.split('\n');
|
||||
|
||||
for (const termConfig of searchTerms) {
|
||||
let regex;
|
||||
let term, searchIn, pattern, description, flags;
|
||||
|
||||
// Handle different input formats (string or object)
|
||||
if (typeof termConfig === 'string') {
|
||||
term = termConfig;
|
||||
searchIn = 'both'; // default
|
||||
} else {
|
||||
term = termConfig.term;
|
||||
searchIn = termConfig.searchIn || 'both';
|
||||
pattern = termConfig.pattern;
|
||||
description = termConfig.description;
|
||||
flags = termConfig.flags;
|
||||
}
|
||||
|
||||
// Skip if this term shouldn't be searched in the current location
|
||||
if (searchIn !== 'both' && searchIn !== searchLocation) {
|
||||
continue;
|
||||
}
|
||||
|
||||
// Create appropriate regex
|
||||
if (searchType === 'regex') {
|
||||
regex = new RegExp(pattern, flags || "gi");
|
||||
} else {
|
||||
regex = createSearchRegex(term, searchType);
|
||||
}
|
||||
|
||||
const termMatches = [];
|
||||
|
||||
// Check each line for matches
|
||||
lines.forEach((line, lineIndex) => {
|
||||
const lineMatches = line.match(regex);
|
||||
if (lineMatches) {
|
||||
lineMatches.forEach(match => {
|
||||
termMatches.push({
|
||||
match: match,
|
||||
lineNumber: lineIndex + 1,
|
||||
lineContent: line.trim(),
|
||||
searchType: searchType,
|
||||
searchLocation: searchLocation,
|
||||
originalTerm: term || pattern,
|
||||
description: description,
|
||||
// Show context around the match in the line
|
||||
context: line.length > 100 ?
|
||||
line.substring(Math.max(0, line.toLowerCase().indexOf(match.toLowerCase()) - 30),
|
||||
line.toLowerCase().indexOf(match.toLowerCase()) + match.length + 30) + '...'
|
||||
: line.trim()
|
||||
});
|
||||
});
|
||||
}
|
||||
});
|
||||
|
||||
if (termMatches.length > 0) {
|
||||
matches.push({
|
||||
term: term || (description || pattern),
|
||||
searchType: searchType,
|
||||
searchLocation: searchLocation,
|
||||
searchIn: searchIn,
|
||||
pattern: pattern,
|
||||
matches: termMatches,
|
||||
count: termMatches.length
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
return matches;
|
||||
}
|
||||
|
||||
// Helper function to check if label should be added
|
||||
async function processLabel(labelName, config) {
|
||||
const body = context.payload.issue.body || "";
|
||||
const title = context.payload.issue.title || "";
|
||||
|
||||
core.notice(`Processing label: ${labelName}`);
|
||||
core.notice(`Issue Title: "${title}"`);
|
||||
core.notice(`Issue Body length: ${body.length} characters`);
|
||||
|
||||
let shouldAddLabel = false;
|
||||
let allMatches = [];
|
||||
let reason = '';
|
||||
|
||||
const keywords = config.keywords || [];
|
||||
const substrings = config.substrings || [];
|
||||
const regexPatterns = config.regexPatterns || [];
|
||||
|
||||
core.notice(`Searching with ${keywords.length} keywords, ${substrings.length} substrings, and ${regexPatterns.length} regex patterns`);
|
||||
|
||||
// Search in title
|
||||
if (title.trim()) {
|
||||
core.notice(`Searching in title: "${title}"`);
|
||||
|
||||
const titleKeywordMatches = findMatchingTermsWithLines(title, keywords, 'keyword', 'title');
|
||||
const titleSubstringMatches = findMatchingTermsWithLines(title, substrings, 'substring', 'title');
|
||||
const titleRegexMatches = findMatchingTermsWithLines(title, regexPatterns, 'regex', 'title');
|
||||
|
||||
allMatches.push(...titleKeywordMatches, ...titleSubstringMatches, ...titleRegexMatches);
|
||||
}
|
||||
|
||||
// Search in body
|
||||
if (body.trim()) {
|
||||
core.notice(`Searching in body (${body.length} characters)`);
|
||||
|
||||
const bodyKeywordMatches = findMatchingTermsWithLines(body, keywords, 'keyword', 'body');
|
||||
const bodySubstringMatches = findMatchingTermsWithLines(body, substrings, 'substring', 'body');
|
||||
const bodyRegexMatches = findMatchingTermsWithLines(body, regexPatterns, 'regex', 'body');
|
||||
|
||||
allMatches.push(...bodyKeywordMatches, ...bodySubstringMatches, ...bodyRegexMatches);
|
||||
}
|
||||
|
||||
if (allMatches.length > 0) {
|
||||
core.notice(`Found ${allMatches.length} matching term(s):`);
|
||||
|
||||
for (const termMatch of allMatches) {
|
||||
const locationText = termMatch.searchLocation === 'title' ? 'title' : 'body';
|
||||
const searchInText = termMatch.searchIn === 'both' ? 'both' : termMatch.searchIn;
|
||||
|
||||
if (termMatch.searchType === 'regex') {
|
||||
core.notice(` 📍 Regex: "${termMatch.term}" (pattern: ${termMatch.pattern}) found ${termMatch.count} time(s) in ${locationText} (configured to search in: ${searchInText}):`);
|
||||
} else {
|
||||
core.notice(` 📍 Term: "${termMatch.term}" (${termMatch.searchType} search) found ${termMatch.count} time(s) in ${locationText} (configured to search in: ${searchInText}):`);
|
||||
}
|
||||
|
||||
// Show details for each match
|
||||
termMatch.matches.forEach((match, index) => {
|
||||
core.notice(` ${index + 1}. Line ${match.lineNumber} in ${match.searchLocation}: "${match.match}" [${match.searchType}]`);
|
||||
if (match.description) {
|
||||
core.notice(` Description: ${match.description}`);
|
||||
}
|
||||
core.notice(` Context: ${match.context}`);
|
||||
if (match.lineContent !== match.context) {
|
||||
core.notice(` Full line: ${match.lineContent}`);
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
shouldAddLabel = true;
|
||||
const totalMatches = allMatches.reduce((sum, t) => sum + t.count, 0);
|
||||
const titleMatches = allMatches.filter(t => t.searchLocation === 'title').reduce((sum, t) => sum + t.count, 0);
|
||||
const bodyMatches = allMatches.filter(t => t.searchLocation === 'body').reduce((sum, t) => sum + t.count, 0);
|
||||
const keywordMatches = allMatches.filter(t => t.searchType === 'keyword').reduce((sum, t) => sum + t.count, 0);
|
||||
const substringMatches = allMatches.filter(t => t.searchType === 'substring').reduce((sum, t) => sum + t.count, 0);
|
||||
const regexMatches = allMatches.filter(t => t.searchType === 'regex').reduce((sum, t) => sum + t.count, 0);
|
||||
|
||||
reason = `Found ${totalMatches} total matches (${titleMatches} in title, ${bodyMatches} in body) - ${keywordMatches} keyword matches, ${substringMatches} substring matches, ${regexMatches} regex matches`;
|
||||
}
|
||||
|
||||
core.notice(`Final decision: ${shouldAddLabel ? 'ADD LABEL' : 'DO NOT ADD LABEL'}`);
|
||||
core.notice(`Reason: ${reason || 'No matching terms found'}`);
|
||||
|
||||
if (shouldAddLabel) {
|
||||
const existingLabels = context.payload.issue.labels.map(l => l.name);
|
||||
if (!existingLabels.includes(labelName)) {
|
||||
await github.rest.issues.addLabels({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
labels: [labelName],
|
||||
});
|
||||
core.notice(`Label "${labelName}" added. ${reason}`);
|
||||
return true;
|
||||
}
|
||||
core.notice(`Label "${labelName}" already present.`);
|
||||
return false;
|
||||
}
|
||||
|
||||
core.notice(`No matching terms found for label "${labelName}".`);
|
||||
return false;
|
||||
}
|
||||
|
||||
// Process all configured labels
|
||||
const processLabels = Object.entries(labelConfig)
|
||||
.map(([labelName, config]) => processLabel(labelName, config));
|
||||
const labelsAdded = await Promise.all(processLabels);
|
||||
const numLabelsAdded = labelsAdded.reduce((x, y) => x + y, 0);
|
||||
core.notice(`Processing complete. ${numLabelsAdded} label(s) added.`);
|
||||
85
.github/workflows/lint-and-deploy.yaml
vendored
Normal file
85
.github/workflows/lint-and-deploy.yaml
vendored
Normal file
@ -0,0 +1,85 @@
|
||||
name: Lint and Deploy Charts
|
||||
|
||||
on: pull_request
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
|
||||
jobs:
|
||||
lint-and-deploy:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Checkout
|
||||
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
with:
|
||||
fetch-depth: 0
|
||||
|
||||
- name: Set up Helm
|
||||
uses: azure/setup-helm@b9e51907a09c216f16ebe8536097933489208112 # v4.3.0
|
||||
with:
|
||||
version: v3.14.4
|
||||
|
||||
#Python is required because ct lint runs Yamale and yamllint which require Python.
|
||||
- uses: actions/setup-python@42375524e23c412d93fb67b49958b491fce71c38 # v5.4.0
|
||||
with:
|
||||
python-version: '3.13'
|
||||
|
||||
- name: Set up chart-testing
|
||||
uses: helm/chart-testing-action@0d28d3144d3a25ea2cc349d6e59901c4ff469b3b # v2.7.0
|
||||
with:
|
||||
version: v3.10.1
|
||||
|
||||
- name: Run chart-testing (lint)
|
||||
run: ct lint --target-branch ${{ github.event.repository.default_branch }} --chart-dirs examples/online_serving/chart-helm --charts examples/online_serving/chart-helm
|
||||
|
||||
- name: Setup minio
|
||||
run: |
|
||||
docker network create vllm-net
|
||||
docker run -d -p 9000:9000 --name minio --net vllm-net \
|
||||
-e "MINIO_ACCESS_KEY=minioadmin" \
|
||||
-e "MINIO_SECRET_KEY=minioadmin" \
|
||||
-v /tmp/data:/data \
|
||||
-v /tmp/config:/root/.minio \
|
||||
minio/minio server /data
|
||||
export AWS_ACCESS_KEY_ID=minioadmin
|
||||
export AWS_SECRET_ACCESS_KEY=minioadmin
|
||||
export AWS_EC2_METADATA_DISABLED=true
|
||||
mkdir opt-125m
|
||||
cd opt-125m && curl -O -Ls "https://huggingface.co/facebook/opt-125m/resolve/main/{pytorch_model.bin,config.json,generation_config.json,merges.txt,special_tokens_map.json,tokenizer_config.json,vocab.json}" && cd ..
|
||||
aws --endpoint-url http://127.0.0.1:9000/ s3 mb s3://testbucket
|
||||
aws --endpoint-url http://127.0.0.1:9000/ s3 cp opt-125m/ s3://testbucket/opt-125m --recursive
|
||||
|
||||
- name: Create kind cluster
|
||||
uses: helm/kind-action@a1b0e391336a6ee6713a0583f8c6240d70863de3 # v1.12.0
|
||||
|
||||
- name: Build the Docker image vllm cpu
|
||||
run: docker buildx build -f docker/Dockerfile.cpu -t vllm-cpu-env .
|
||||
|
||||
- name: Configuration of docker images, network and namespace for the kind cluster
|
||||
run: |
|
||||
docker pull amazon/aws-cli:2.6.4
|
||||
kind load docker-image amazon/aws-cli:2.6.4 --name chart-testing
|
||||
kind load docker-image vllm-cpu-env:latest --name chart-testing
|
||||
docker network connect vllm-net "$(docker ps -aqf "name=chart-testing-control-plane")"
|
||||
kubectl create ns ns-vllm
|
||||
|
||||
- name: Run chart-testing (install)
|
||||
run: |
|
||||
export AWS_ACCESS_KEY_ID=minioadmin
|
||||
export AWS_SECRET_ACCESS_KEY=minioadmin
|
||||
sleep 30 && kubectl -n ns-vllm logs -f "$(kubectl -n ns-vllm get pods | awk '/deployment/ {print $1;exit}')" &
|
||||
helm install --wait --wait-for-jobs --timeout 5m0s --debug --create-namespace --namespace=ns-vllm test-vllm examples/online_serving/chart-helm -f examples/online_serving/chart-helm/values.yaml --set secrets.s3endpoint=http://minio:9000 --set secrets.s3bucketname=testbucket --set secrets.s3accesskeyid=$AWS_ACCESS_KEY_ID --set secrets.s3accesskey=$AWS_SECRET_ACCESS_KEY --set resources.requests.cpu=1 --set resources.requests.memory=4Gi --set resources.limits.cpu=2 --set resources.limits.memory=5Gi --set image.env[0].name=VLLM_CPU_KVCACHE_SPACE --set image.env[1].name=VLLM_LOGGING_LEVEL --set-string image.env[0].value="1" --set-string image.env[1].value="DEBUG" --set-string extraInit.s3modelpath="opt-125m/" --set-string 'resources.limits.nvidia\.com/gpu=0' --set-string 'resources.requests.nvidia\.com/gpu=0' --set-string image.repository="vllm-cpu-env"
|
||||
|
||||
- name: curl test
|
||||
run: |
|
||||
kubectl -n ns-vllm port-forward service/test-vllm-service 8001:80 &
|
||||
sleep 10
|
||||
CODE="$(curl -v -f --location http://localhost:8001/v1/completions \
|
||||
--header "Content-Type: application/json" \
|
||||
--data '{
|
||||
"model": "opt-125m",
|
||||
"prompt": "San Francisco is a",
|
||||
"max_tokens": 7,
|
||||
"temperature": 0
|
||||
}'):$CODE"
|
||||
echo "$CODE"
|
||||
17
.github/workflows/matchers/markdownlint.json
vendored
17
.github/workflows/matchers/markdownlint.json
vendored
@ -1,17 +0,0 @@
|
||||
{
|
||||
"problemMatcher": [
|
||||
{
|
||||
"owner": "markdownlint",
|
||||
"pattern": [
|
||||
{
|
||||
"regexp": "^([^:]*):(\\d+):?(\\d+)?\\s([\\w-\\/]*)\\s(.*)$",
|
||||
"file": 1,
|
||||
"line": 2,
|
||||
"column": 3,
|
||||
"code": 4,
|
||||
"message": 5
|
||||
}
|
||||
]
|
||||
}
|
||||
]
|
||||
}
|
||||
5
.github/workflows/pre-commit.yml
vendored
5
.github/workflows/pre-commit.yml
vendored
@ -5,10 +5,6 @@ on:
|
||||
push:
|
||||
branches: [main]
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.ref }}
|
||||
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
|
||||
@ -21,7 +17,6 @@ jobs:
|
||||
with:
|
||||
python-version: "3.12"
|
||||
- run: echo "::add-matcher::.github/workflows/matchers/actionlint.json"
|
||||
- run: echo "::add-matcher::.github/workflows/matchers/markdownlint.json"
|
||||
- run: echo "::add-matcher::.github/workflows/matchers/mypy.json"
|
||||
- uses: pre-commit/action@2c7b3805fd2a0fd8c1884dcaebf91fc102a13ecd # v3.0.1
|
||||
with:
|
||||
|
||||
111
.github/workflows/publish.yml
vendored
Normal file
111
.github/workflows/publish.yml
vendored
Normal file
@ -0,0 +1,111 @@
|
||||
# This workflow will upload a Python Package to Release asset
|
||||
# For more information see: https://help.github.com/en/actions/language-and-framework-guides/using-python-with-github-actions
|
||||
|
||||
name: Create Release
|
||||
|
||||
on:
|
||||
push:
|
||||
tags:
|
||||
- v*
|
||||
|
||||
# Needed to create release and upload assets
|
||||
permissions:
|
||||
contents: write
|
||||
|
||||
jobs:
|
||||
release:
|
||||
# Retrieve tag and create release
|
||||
name: Create Release
|
||||
runs-on: ubuntu-latest
|
||||
outputs:
|
||||
upload_url: ${{ steps.create_release.outputs.upload_url }}
|
||||
steps:
|
||||
- name: Checkout
|
||||
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
|
||||
- name: Extract branch info
|
||||
shell: bash
|
||||
run: |
|
||||
echo "release_tag=${GITHUB_REF#refs/*/}" >> "$GITHUB_ENV"
|
||||
|
||||
- name: Create Release
|
||||
id: create_release
|
||||
uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
|
||||
env:
|
||||
RELEASE_TAG: ${{ env.release_tag }}
|
||||
with:
|
||||
github-token: "${{ secrets.GITHUB_TOKEN }}"
|
||||
script: |
|
||||
const script = require('.github/workflows/scripts/create_release.js')
|
||||
await script(github, context, core)
|
||||
|
||||
# NOTE(simon): No longer build wheel using GitHub Actions. See buildkite's release workflow.
|
||||
# wheel:
|
||||
# name: Build Wheel
|
||||
# runs-on: ${{ matrix.os }}
|
||||
# needs: release
|
||||
|
||||
# strategy:
|
||||
# fail-fast: false
|
||||
# matrix:
|
||||
# os: ['ubuntu-20.04']
|
||||
# python-version: ['3.9', '3.10', '3.11', '3.12']
|
||||
# pytorch-version: ['2.4.0'] # Must be the most recent version that meets requirements/cuda.txt.
|
||||
# cuda-version: ['11.8', '12.1']
|
||||
|
||||
# steps:
|
||||
# - name: Checkout
|
||||
# uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
|
||||
# - name: Setup ccache
|
||||
# uses: hendrikmuhs/ccache-action@ed74d11c0b343532753ecead8a951bb09bb34bc9 # v1.2.14
|
||||
# with:
|
||||
# create-symlink: true
|
||||
# key: ${{ github.job }}-${{ matrix.python-version }}-${{ matrix.cuda-version }}
|
||||
|
||||
# - name: Set up Linux Env
|
||||
# if: ${{ runner.os == 'Linux' }}
|
||||
# run: |
|
||||
# bash -x .github/workflows/scripts/env.sh
|
||||
|
||||
# - name: Set up Python
|
||||
# uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # v5.3.0
|
||||
# with:
|
||||
# python-version: ${{ matrix.python-version }}
|
||||
|
||||
# - name: Install CUDA ${{ matrix.cuda-version }}
|
||||
# run: |
|
||||
# bash -x .github/workflows/scripts/cuda-install.sh ${{ matrix.cuda-version }} ${{ matrix.os }}
|
||||
|
||||
# - name: Install PyTorch ${{ matrix.pytorch-version }} with CUDA ${{ matrix.cuda-version }}
|
||||
# run: |
|
||||
# bash -x .github/workflows/scripts/pytorch-install.sh ${{ matrix.python-version }} ${{ matrix.pytorch-version }} ${{ matrix.cuda-version }}
|
||||
|
||||
# - name: Build wheel
|
||||
# shell: bash
|
||||
# env:
|
||||
# CMAKE_BUILD_TYPE: Release # do not compile with debug symbol to reduce wheel size
|
||||
# run: |
|
||||
# bash -x .github/workflows/scripts/build.sh ${{ matrix.python-version }} ${{ matrix.cuda-version }}
|
||||
# wheel_name=$(find dist -name "*whl" -print0 | xargs -0 -n 1 basename)
|
||||
# asset_name=${wheel_name//"linux"/"manylinux1"}
|
||||
# echo "wheel_name=${wheel_name}" >> "$GITHUB_ENV"
|
||||
# echo "asset_name=${asset_name}" >> "$GITHUB_ENV"
|
||||
|
||||
# - name: Upload Release Asset
|
||||
# uses: actions/upload-release-asset@e8f9f06c4b078e705bd2ea027f0926603fc9b4d5 # v1.0.2
|
||||
# env:
|
||||
# GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
|
||||
# with:
|
||||
# upload_url: ${{ needs.release.outputs.upload_url }}
|
||||
# asset_path: ./dist/${{ env.wheel_name }}
|
||||
# asset_name: ${{ env.asset_name }}
|
||||
# asset_content_type: application/*
|
||||
|
||||
# (Danielkinz): This last step will publish the .whl to pypi. Warning: untested
|
||||
# - name: Publish package
|
||||
# uses: pypa/gh-action-pypi-publish@release/v1.8
|
||||
# with:
|
||||
# repository-url: https://test.pypi.org/legacy/
|
||||
# password: ${{ secrets.PYPI_API_TOKEN }}
|
||||
# skip-existing: true
|
||||
49
.github/workflows/reminder_comment.yml
vendored
49
.github/workflows/reminder_comment.yml
vendored
@ -12,43 +12,16 @@ jobs:
|
||||
uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
|
||||
with:
|
||||
script: |
|
||||
try {
|
||||
// Get the PR author
|
||||
const prAuthor = context.payload.pull_request.user.login;
|
||||
|
||||
// Check if this is the author's first PR in this repository
|
||||
// Use GitHub's search API to find all PRs by this author
|
||||
const { data: searchResults } = await github.rest.search.issuesAndPullRequests({
|
||||
q: `repo:${context.repo.owner}/${context.repo.repo} type:pr author:${prAuthor}`,
|
||||
per_page: 100
|
||||
});
|
||||
|
||||
const authorPRCount = searchResults.total_count;
|
||||
|
||||
console.log(`Found ${authorPRCount} PRs by ${prAuthor}`);
|
||||
|
||||
// Only post comment if this is the first PR (only one PR by this author)
|
||||
if (authorPRCount === 1) {
|
||||
console.log(`Posting welcome comment for first-time contributor: ${prAuthor}`);
|
||||
await github.rest.issues.createComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
body: '👋 Hi! Thank you for contributing to the vLLM project.\n\n' +
|
||||
'💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.\n\n' +
|
||||
'Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run `fastcheck` CI which starts running only a small and essential subset of CI tests to quickly catch errors. \n\n' +
|
||||
'You ask your reviewers to trigger select CI tests on top of `fastcheck` CI. \n\n' +
|
||||
'Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.\n\n' +
|
||||
'To run CI, PR reviewers can either: Add `ready` label to the PR or enable auto-merge.\n\n' +
|
||||
'If you have any questions, please reach out to us on Slack at https://slack.vllm.ai.\n\n' +
|
||||
'🚀'
|
||||
});
|
||||
} else {
|
||||
console.log(`Skipping comment for ${prAuthor} - not their first PR (${authorPRCount} PRs found)`);
|
||||
}
|
||||
} catch (error) {
|
||||
console.error('Error checking PR history or posting comment:', error);
|
||||
// Don't fail the workflow, just log the error
|
||||
}
|
||||
github.rest.issues.createComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
body: '👋 Hi! Thank you for contributing to the vLLM project.\n\n' +
|
||||
'💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.\n\n' +
|
||||
'Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run `fastcheck` CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your `fastcheck` build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping `simon-mo` or `khluu` to add you in our Buildkite org.\n\n' +
|
||||
'Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.\n\n' +
|
||||
'To run CI, PR reviewers can either: Add `ready` label to the PR or enable auto-merge.\n\n' +
|
||||
'🚀'
|
||||
})
|
||||
env:
|
||||
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
|
||||
|
||||
1
.github/workflows/scripts/build.sh
vendored
1
.github/workflows/scripts/build.sh
vendored
@ -15,6 +15,7 @@ $python_executable -m pip install -r requirements/build.txt -r requirements/cuda
|
||||
export MAX_JOBS=1
|
||||
# Make sure release wheels are built for the following architectures
|
||||
export TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 8.9 9.0+PTX"
|
||||
export VLLM_FA_CMAKE_GPU_ARCHES="80-real;90-real"
|
||||
|
||||
bash tools/check_repo.sh
|
||||
|
||||
|
||||
10
.gitignore
vendored
10
.gitignore
vendored
@ -4,9 +4,6 @@
|
||||
# vllm-flash-attn built from source
|
||||
vllm/vllm_flash_attn/*
|
||||
|
||||
# triton jit
|
||||
.triton
|
||||
|
||||
# Byte-compiled / optimized / DLL files
|
||||
__pycache__/
|
||||
*.py[cod]
|
||||
@ -149,9 +146,7 @@ venv.bak/
|
||||
|
||||
# mkdocs documentation
|
||||
/site
|
||||
docs/argparse
|
||||
docs/examples/*
|
||||
!docs/examples/README.md
|
||||
docs/examples
|
||||
|
||||
# mypy
|
||||
.mypy_cache/
|
||||
@ -207,6 +202,3 @@ shellcheck*/
|
||||
|
||||
# Ignore moe/marlin_moe gen code
|
||||
csrc/moe/marlin_moe_wna16/kernel_*
|
||||
|
||||
# Ignore ep_kernels_workspace folder
|
||||
ep_kernels_workspace/
|
||||
@ -1,13 +0,0 @@
|
||||
MD007:
|
||||
indent: 4
|
||||
MD013: false
|
||||
MD024:
|
||||
siblings_only: true
|
||||
MD033: false
|
||||
MD042: false
|
||||
MD045: false
|
||||
MD046: false
|
||||
MD051: false
|
||||
MD052: false
|
||||
MD053: false
|
||||
MD059: false
|
||||
@ -21,7 +21,7 @@ repos:
|
||||
- id: ruff-format
|
||||
files: ^(.buildkite|benchmarks|examples)/.*
|
||||
- repo: https://github.com/crate-ci/typos
|
||||
rev: v1.35.5
|
||||
rev: v1.32.0
|
||||
hooks:
|
||||
- id: typos
|
||||
- repo: https://github.com/PyCQA/isort
|
||||
@ -35,12 +35,12 @@ repos:
|
||||
exclude: 'csrc/(moe/topk_softmax_kernels.cu|quantization/gguf/(ggml-common.h|dequantize.cuh|vecdotq.cuh|mmq.cuh|mmvq.cuh))|vllm/third_party/.*'
|
||||
types_or: [c++, cuda]
|
||||
args: [--style=file, --verbose]
|
||||
- repo: https://github.com/igorshubovych/markdownlint-cli
|
||||
rev: v0.45.0
|
||||
- repo: https://github.com/jackdewinter/pymarkdown
|
||||
rev: v0.9.29
|
||||
hooks:
|
||||
- id: markdownlint
|
||||
- id: pymarkdown
|
||||
exclude: '.*\.inc\.md'
|
||||
stages: [manual] # Only run in CI
|
||||
args: [fix]
|
||||
- repo: https://github.com/rhysd/actionlint
|
||||
rev: v1.7.7
|
||||
hooks:
|
||||
@ -166,11 +166,11 @@ repos:
|
||||
language: python
|
||||
types: [python]
|
||||
pass_filenames: true
|
||||
files: vllm/config.py|tests/test_config.py|vllm/entrypoints/openai/cli_args.py
|
||||
files: vllm/config.py|tests/test_config.py
|
||||
# Keep `suggestion` last
|
||||
- id: suggestion
|
||||
name: Suggestion
|
||||
entry: bash -c 'echo "To bypass all the pre-commit hooks, add --no-verify to git commit. To skip a specific hook, prefix the commit command with SKIP=<hook-id>."'
|
||||
entry: bash -c 'echo "To bypass pre-commit hooks, add --no-verify to git commit."'
|
||||
language: system
|
||||
verbose: true
|
||||
pass_filenames: false
|
||||
|
||||
@ -7,9 +7,6 @@ build:
|
||||
os: ubuntu-22.04
|
||||
tools:
|
||||
python: "3.12"
|
||||
jobs:
|
||||
post_checkout:
|
||||
- git fetch --unshallow || true
|
||||
|
||||
mkdocs:
|
||||
configuration: mkdocs.yaml
|
||||
|
||||
173
CMakeLists.txt
173
CMakeLists.txt
@ -30,7 +30,7 @@ install(CODE "set(CMAKE_INSTALL_LOCAL_ONLY TRUE)" ALL_COMPONENTS)
|
||||
# Supported python versions. These versions will be searched in order, the
|
||||
# first match will be selected. These should be kept in sync with setup.py.
|
||||
#
|
||||
set(PYTHON_SUPPORTED_VERSIONS "3.9" "3.10" "3.11" "3.12" "3.13")
|
||||
set(PYTHON_SUPPORTED_VERSIONS "3.9" "3.10" "3.11" "3.12")
|
||||
|
||||
# Supported AMD GPU architectures.
|
||||
set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1030;gfx1100;gfx1101;gfx1200;gfx1201")
|
||||
@ -45,8 +45,8 @@ set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1030;gfx1100;gfx1
|
||||
# requirements.txt files and should be kept consistent. The ROCm torch
|
||||
# versions are derived from docker/Dockerfile.rocm
|
||||
#
|
||||
set(TORCH_SUPPORTED_VERSION_CUDA "2.8.0")
|
||||
set(TORCH_SUPPORTED_VERSION_ROCM "2.8.0")
|
||||
set(TORCH_SUPPORTED_VERSION_CUDA "2.7.0")
|
||||
set(TORCH_SUPPORTED_VERSION_ROCM "2.7.0")
|
||||
|
||||
#
|
||||
# Try to find python package with an executable that exactly matches
|
||||
@ -171,6 +171,7 @@ if(NVCC_THREADS AND VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND VLLM_GPU_FLAGS "--threads=${NVCC_THREADS}")
|
||||
endif()
|
||||
|
||||
|
||||
#
|
||||
# Use FetchContent for C++ dependencies that are compiled as part of vLLM's build process.
|
||||
# setup.py will override FETCHCONTENT_BASE_DIR to play nicely with sccache.
|
||||
@ -231,6 +232,7 @@ endif()
|
||||
|
||||
set(VLLM_EXT_SRC
|
||||
"csrc/mamba/mamba_ssm/selective_scan_fwd.cu"
|
||||
"csrc/mamba/causal_conv1d/causal_conv1d.cu"
|
||||
"csrc/cache_kernels.cu"
|
||||
"csrc/attention/paged_attention_v1.cu"
|
||||
"csrc/attention/paged_attention_v2.cu"
|
||||
@ -249,6 +251,7 @@ set(VLLM_EXT_SRC
|
||||
"csrc/quantization/gguf/gguf_kernel.cu"
|
||||
"csrc/quantization/activation_kernels.cu"
|
||||
"csrc/cuda_utils_kernels.cu"
|
||||
"csrc/prepare_inputs/advance_step.cu"
|
||||
"csrc/custom_all_reduce.cu"
|
||||
"csrc/torch_bindings.cpp")
|
||||
|
||||
@ -256,7 +259,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
SET(CUTLASS_ENABLE_HEADERS_ONLY ON CACHE BOOL "Enable only the header library")
|
||||
|
||||
# Set CUTLASS_REVISION. Used for FetchContent. Also fixes some bogus messages when building.
|
||||
set(CUTLASS_REVISION "v4.0.0" CACHE STRING "CUTLASS revision to use")
|
||||
set(CUTLASS_REVISION "v3.9.2" CACHE STRING "CUTLASS revision to use")
|
||||
|
||||
# Use the specified CUTLASS source directory for compilation if VLLM_CUTLASS_SRC_DIR is provided
|
||||
if (DEFINED ENV{VLLM_CUTLASS_SRC_DIR})
|
||||
@ -286,6 +289,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
FetchContent_MakeAvailable(cutlass)
|
||||
|
||||
list(APPEND VLLM_EXT_SRC
|
||||
"csrc/quantization/aqlm/gemm_kernels.cu"
|
||||
"csrc/quantization/awq/gemm_kernels.cu"
|
||||
"csrc/permute_cols.cu"
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_entry.cu"
|
||||
@ -294,8 +298,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
"csrc/quantization/fp4/nvfp4_blockwise_moe_kernel.cu"
|
||||
"csrc/sparse/cutlass/sparse_scaled_mm_entry.cu"
|
||||
"csrc/cutlass_extensions/common.cpp"
|
||||
"csrc/attention/mla/cutlass_mla_entry.cu"
|
||||
"csrc/quantization/fp8/per_token_group_quant.cu")
|
||||
"csrc/attention/mla/cutlass_mla_entry.cu")
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${VLLM_EXT_SRC}"
|
||||
@ -349,27 +352,20 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${MARLIN_TEMPLATE_KERNEL_SRC}"
|
||||
CUDA_ARCHS "${MARLIN_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8)
|
||||
set_source_files_properties(${MARLIN_TEMPLATE_KERNEL_SRC}
|
||||
PROPERTIES COMPILE_FLAGS "-static-global-template-stub=false")
|
||||
endif()
|
||||
|
||||
list(APPEND VLLM_EXT_SRC ${MARLIN_TEMPLATE_KERNEL_SRC})
|
||||
|
||||
set(MARLIN_SRCS
|
||||
"csrc/quantization/marlin/dense/marlin_cuda_kernel.cu"
|
||||
"csrc/quantization/marlin/sparse/marlin_24_cuda_kernel.cu"
|
||||
"csrc/quantization/marlin/qqq/marlin_qqq_gemm_kernel.cu"
|
||||
"csrc/quantization/gptq_marlin/gptq_marlin.cu"
|
||||
"csrc/quantization/gptq_marlin/gptq_marlin_repack.cu"
|
||||
"csrc/quantization/gptq_marlin/awq_marlin_repack.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${MARLIN_SRCS}"
|
||||
CUDA_ARCHS "${MARLIN_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8)
|
||||
set_source_files_properties("csrc/quantization/gptq_marlin/gptq_marlin.cu"
|
||||
PROPERTIES COMPILE_FLAGS "-static-global-template-stub=false")
|
||||
endif()
|
||||
list(APPEND VLLM_EXT_SRC "${MARLIN_SRCS}")
|
||||
|
||||
message(STATUS "Building Marlin kernels for archs: ${MARLIN_ARCHS}")
|
||||
else()
|
||||
message(STATUS "Not building Marlin kernels as no compatible archs found"
|
||||
@ -397,7 +393,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The cutlass_scaled_mm kernels for Hopper (c3x, i.e. CUTLASS 3.x) require
|
||||
# CUDA 12.0 or later
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a;" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0 AND SCALED_MM_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0 AND SCALED_MM_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm90.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm90_fp8.cu"
|
||||
@ -413,7 +409,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND SCALED_MM_3X_ARCHS "${SCALED_MM_ARCHS}")
|
||||
message(STATUS "Building scaled_mm_c3x_sm90 for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0 AND SCALED_MM_ARCHS)
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building scaled_mm_c3x_sm90 as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
"later if you intend on running FP8 quantized models on "
|
||||
@ -428,11 +424,10 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The cutlass_scaled_mm kernels for Geforce Blackwell SM120 (c3x, i.e. CUTLASS 3.x) require
|
||||
# CUDA 12.8 or later
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0;12.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm120.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm120_fp8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_blockwise_sm120_fp8.cu"
|
||||
)
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
@ -443,7 +438,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND SCALED_MM_3X_ARCHS "${SCALED_MM_ARCHS}")
|
||||
message(STATUS "Building scaled_mm_c3x_sm120 for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building scaled_mm_c3x_sm120 as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or "
|
||||
"later if you intend on running FP8 quantized models on "
|
||||
@ -458,7 +453,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The cutlass_scaled_mm kernels for Blackwell SM100 (c3x, i.e. CUTLASS 3.x)
|
||||
# require CUDA 12.8 or later
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm100.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm100_fp8.cu"
|
||||
@ -473,7 +468,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND SCALED_MM_3X_ARCHS "${SCALED_MM_ARCHS}")
|
||||
message(STATUS "Building scaled_mm_c3x_sm100 for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building scaled_mm_c3x_sm100 as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or "
|
||||
"later if you intend on running FP8 quantized models on "
|
||||
@ -516,7 +511,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The 2:4 sparse kernels cutlass_scaled_sparse_mm and cutlass_compressor
|
||||
# require CUDA 12.2 or later (and only work on Hopper).
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a;" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.2 AND SCALED_MM_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.2 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/sparse/cutlass/sparse_scaled_mm_c3x.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
@ -525,7 +520,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_SPARSE_SCALED_MM_C3X=1")
|
||||
message(STATUS "Building sparse_scaled_mm_c3x for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.2 AND SCALED_MM_ARCHS)
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.2 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building sparse_scaled_mm_c3x kernels as CUDA Compiler version is "
|
||||
"not >= 12.2, we recommend upgrading to CUDA 12.2 or later "
|
||||
"if you intend on running FP8 sparse quantized models on Hopper.")
|
||||
@ -535,32 +530,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# The nvfp4_scaled_mm_sm120 kernels for Geforce Blackwell SM120 require
|
||||
# CUDA 12.8 or later
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "12.0;12.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND FP4_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/fp4/nvfp4_quant_kernels.cu"
|
||||
"csrc/quantization/fp4/activation_nvfp4_quant_fusion_kernels.cu"
|
||||
"csrc/quantization/fp4/nvfp4_scaled_mm_sm120_kernels.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${FP4_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_NVFP4_SM120=1")
|
||||
message(STATUS "Building NVFP4 for archs: ${FP4_ARCHS}")
|
||||
else()
|
||||
message(STATUS "Not building NVFP4 as no compatible archs were found.")
|
||||
# clear FP4_ARCHS
|
||||
set(FP4_ARCHS)
|
||||
endif()
|
||||
|
||||
# FP4 Archs and flags
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND FP4_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND FP4_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/fp4/nvfp4_quant_kernels.cu"
|
||||
"csrc/quantization/fp4/activation_nvfp4_quant_fusion_kernels.cu"
|
||||
"csrc/quantization/fp4/nvfp4_experts_quant.cu"
|
||||
"csrc/quantization/fp4/nvfp4_scaled_mm_kernels.cu"
|
||||
"csrc/quantization/fp4/nvfp4_blockwise_moe_kernel.cu")
|
||||
@ -568,7 +542,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${FP4_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_NVFP4_SM100=1")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_NVFP4=1")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_CUTLASS_MOE_SM100=1")
|
||||
message(STATUS "Building NVFP4 for archs: ${FP4_ARCHS}")
|
||||
else()
|
||||
@ -579,10 +553,9 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
|
||||
# CUTLASS MLA Archs and flags
|
||||
cuda_archs_loose_intersection(MLA_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND MLA_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND MLA_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/attention/mla/cutlass_mla_kernels.cu"
|
||||
"csrc/attention/mla/sm100_cutlass_mla_kernel.cu")
|
||||
"csrc/attention/mla/cutlass_mla_kernels.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${MLA_ARCHS}")
|
||||
@ -605,7 +578,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# if it's possible to compile MoE kernels that use its output.
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/moe/grouped_mm_c3x_sm90.cu")
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/moe/grouped_mm_c3x.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
@ -623,26 +596,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/moe/grouped_mm_c3x_sm100.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_CUTLASS_MOE_SM100=1")
|
||||
message(STATUS "Building grouped_mm_c3x for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building grouped_mm_c3x kernels as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or later "
|
||||
"if you intend on running FP8 quantized MoE models on Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building grouped_mm_c3x as no compatible archs found "
|
||||
"in CUDA target architectures.")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# moe_data.cu is used by all CUTLASS MoE kernels.
|
||||
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND CUTLASS_MOE_DATA_ARCHS)
|
||||
@ -663,33 +616,13 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/moe/blockwise_scaled_group_mm_sm100.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_CUTLASS_MOE_SM100=1")
|
||||
message(STATUS "Building blockwise_scaled_group_mm_sm100 for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building blockwise_scaled_group_mm_sm100 kernels as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or later "
|
||||
"if you intend on running FP8 quantized MoE models on Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building blockwise_scaled_group_mm_sm100 as no compatible archs found "
|
||||
"in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
#
|
||||
# Machete kernels
|
||||
|
||||
# The machete kernels only work on hopper and require CUDA 12.0 or later.
|
||||
# Only build Machete kernels if we are building for something compatible with sm90a
|
||||
cuda_archs_loose_intersection(MACHETE_ARCHS "9.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0 AND MACHETE_ARCHS)
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0 AND MACHETE_ARCHS)
|
||||
#
|
||||
# For the Machete kernels we automatically generate sources for various
|
||||
# preselected input type pairs and schedules.
|
||||
@ -741,7 +674,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
|
||||
message(STATUS "Building Machete kernels for archs: ${MACHETE_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0
|
||||
AND MACHETE_ARCHS)
|
||||
message(STATUS "Not building Machete kernels as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
@ -752,33 +685,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
"found in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# Only build W4A8 kernels if we are building for something compatible with sm90a
|
||||
cuda_archs_loose_intersection(W4A8_ARCHS "9.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0 AND W4A8_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/cutlass_w4a8/w4a8_mm_entry.cu")
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${W4A8_ARCHS}")
|
||||
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
|
||||
message(STATUS "Building W4A8 kernels for archs: ${W4A8_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0
|
||||
AND W4A8_ARCHS)
|
||||
message(STATUS "Not building W4A8 kernels as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
"later if you intend on running w4a16 quantized models on "
|
||||
"Hopper.")
|
||||
else()
|
||||
message(STATUS "Not building W4A8 kernels as no compatible archs "
|
||||
"found in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# if CUDA endif
|
||||
endif()
|
||||
|
||||
@ -819,17 +725,7 @@ set(VLLM_MOE_EXT_SRC
|
||||
"csrc/moe/topk_softmax_kernels.cu")
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND VLLM_MOE_EXT_SRC
|
||||
"csrc/moe/moe_wna16.cu"
|
||||
"csrc/moe/grouped_topk_kernels.cu")
|
||||
endif()
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set(MOE_PERMUTE_SRC
|
||||
"csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
|
||||
"csrc/moe/moe_permute_unpermute_op.cu")
|
||||
|
||||
list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
|
||||
list(APPEND VLLM_MOE_EXT_SRC "csrc/moe/moe_wna16.cu")
|
||||
endif()
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
@ -890,10 +786,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${MOE_WNAA16_MARLIN_SRC}"
|
||||
CUDA_ARCHS "${MARLIN_MOE_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8)
|
||||
set_source_files_properties(${MOE_WNAA16_MARLIN_SRC}
|
||||
PROPERTIES COMPILE_FLAGS "-static-global-template-stub=false")
|
||||
endif()
|
||||
|
||||
list(APPEND VLLM_MOE_EXT_SRC ${MOE_WNAA16_MARLIN_SRC})
|
||||
|
||||
@ -904,6 +796,17 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set(MOE_PERMUTE_SRC
|
||||
"csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
|
||||
"csrc/moe/moe_permute_unpermute_op.cu")
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${MARLIN_PERMUTE_SRC}"
|
||||
CUDA_ARCHS "${MOE_PERMUTE_ARCHS}")
|
||||
|
||||
list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
|
||||
endif()
|
||||
message(STATUS "Enabling moe extension.")
|
||||
define_gpu_extension_target(
|
||||
_moe_C
|
||||
|
||||
22
README.md
22
README.md
@ -1,4 +1,3 @@
|
||||
<!-- markdownlint-disable MD001 MD041 -->
|
||||
<p align="center">
|
||||
<picture>
|
||||
<source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/vllm-project/vllm/main/docs/assets/logos/vllm-logo-text-dark.png">
|
||||
@ -17,19 +16,14 @@ Easy, fast, and cheap LLM serving for everyone
|
||||
---
|
||||
|
||||
*Latest News* 🔥
|
||||
|
||||
- [2025/08] We hosted [vLLM Singapore Meetup](https://www.sginnovate.com/event/vllm-sg-meet). We shared V1 updates, disaggregated serving and MLLM speedups with speakers from Embedded LLM, AMD, WekaIO, and A*STAR. Please find the meetup slides [here](https://drive.google.com/drive/folders/1ncf3GyqLdqFaB6IeB834E5TZJPLAOiXZ?usp=sharing).
|
||||
- [2025/08] We hosted [vLLM Shanghai Meetup](https://mp.weixin.qq.com/s/pDmAXHcN7Iqc8sUKgJgGtg) focusing on building, developing, and integrating with vLLM! Please find the meetup slides [here](https://drive.google.com/drive/folders/1OvLx39wnCGy_WKq8SiVKf7YcxxYI3WCH).
|
||||
- [2025/08] We hosted [vLLM Korea Meetup](https://luma.com/cgcgprmh) with Red Hat and Rebellions! We shared the latest advancements in vLLM along with project spotlights from the vLLM Korea community. Please find the meetup slides [here](https://drive.google.com/file/d/1bcrrAE1rxUgx0mjIeOWT6hNe2RefC5Hm/view).
|
||||
- [2025/08] We hosted [vLLM Beijing Meetup](https://mp.weixin.qq.com/s/dgkWg1WFpWGO2jCdTqQHxA) focusing on large-scale LLM deployment! Please find the meetup slides [here](https://drive.google.com/drive/folders/1Pid6NSFLU43DZRi0EaTcPgXsAzDvbBqF) and the recording [here](https://www.chaspark.com/#/live/1166916873711665152).
|
||||
- [2025/05] We hosted [NYC vLLM Meetup](https://lu.ma/c1rqyf1f)! Please find the meetup slides [here](https://docs.google.com/presentation/d/1_q_aW_ioMJWUImf1s1YM-ZhjXz8cUeL0IJvaquOYBeA/edit?usp=sharing).
|
||||
- [2025/05] vLLM is now a hosted project under PyTorch Foundation! Please find the announcement [here](https://pytorch.org/blog/pytorch-foundation-welcomes-vllm/).
|
||||
- [2025/04] We hosted [Asia Developer Day](https://www.sginnovate.com/event/limited-availability-morning-evening-slots-remaining-inaugural-vllm-asia-developer-day)! Please find the meetup slides from the vLLM team [here](https://docs.google.com/presentation/d/19cp6Qu8u48ihB91A064XfaXruNYiBOUKrBxAmDOllOo/edit?usp=sharing).
|
||||
- [2025/01] We are excited to announce the alpha release of vLLM V1: A major architectural upgrade with 1.7x speedup! Clean code, optimized execution loop, zero-overhead prefix caching, enhanced multimodal support, and more. Please check out our blog post [here](https://blog.vllm.ai/2025/01/27/v1-alpha-release.html).
|
||||
|
||||
<details>
|
||||
<summary>Previous News</summary>
|
||||
|
||||
- [2025/05] We hosted [NYC vLLM Meetup](https://lu.ma/c1rqyf1f)! Please find the meetup slides [here](https://docs.google.com/presentation/d/1_q_aW_ioMJWUImf1s1YM-ZhjXz8cUeL0IJvaquOYBeA/edit?usp=sharing).
|
||||
- [2025/04] We hosted [Asia Developer Day](https://www.sginnovate.com/event/limited-availability-morning-evening-slots-remaining-inaugural-vllm-asia-developer-day)! Please find the meetup slides from the vLLM team [here](https://docs.google.com/presentation/d/19cp6Qu8u48ihB91A064XfaXruNYiBOUKrBxAmDOllOo/edit?usp=sharing).
|
||||
- [2025/03] We hosted [vLLM x Ollama Inference Night](https://lu.ma/vllm-ollama)! Please find the meetup slides from the vLLM team [here](https://docs.google.com/presentation/d/16T2PDD1YwRnZ4Tu8Q5r6n53c5Lr5c73UV9Vd2_eBo4U/edit?usp=sharing).
|
||||
- [2025/03] We hosted [the first vLLM China Meetup](https://mp.weixin.qq.com/s/n77GibL2corAtQHtVEAzfg)! Please find the meetup slides from vLLM team [here](https://docs.google.com/presentation/d/1REHvfQMKGnvz6p3Fd23HhSO4c8j5WPGZV0bKYLwnHyQ/edit?usp=sharing).
|
||||
- [2025/03] We hosted [the East Coast vLLM Meetup](https://lu.ma/7mu4k4xx)! Please find the meetup slides [here](https://docs.google.com/presentation/d/1NHiv8EUFF1NLd3fEYODm56nDmL26lEeXCaDgyDlTsRs/edit#slide=id.g31441846c39_0_0).
|
||||
@ -52,7 +46,6 @@ Easy, fast, and cheap LLM serving for everyone
|
||||
</details>
|
||||
|
||||
---
|
||||
|
||||
## About
|
||||
|
||||
vLLM is a fast and easy-to-use library for LLM inference and serving.
|
||||
@ -70,11 +63,13 @@ vLLM is fast with:
|
||||
- Speculative decoding
|
||||
- Chunked prefill
|
||||
|
||||
**Performance benchmark**: We include a performance benchmark at the end of [our blog post](https://blog.vllm.ai/2024/09/05/perf-update.html). It compares the performance of vLLM against other LLM serving engines ([TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM), [SGLang](https://github.com/sgl-project/sglang) and [LMDeploy](https://github.com/InternLM/lmdeploy)). The implementation is under [nightly-benchmarks folder](.buildkite/nightly-benchmarks/) and you can [reproduce](https://github.com/vllm-project/vllm/issues/8176) this benchmark using our one-click runnable script.
|
||||
|
||||
vLLM is flexible and easy to use with:
|
||||
|
||||
- Seamless integration with popular Hugging Face models
|
||||
- High-throughput serving with various decoding algorithms, including *parallel sampling*, *beam search*, and more
|
||||
- Tensor, pipeline, data and expert parallelism support for distributed inference
|
||||
- Tensor parallelism and pipeline parallelism support for distributed inference
|
||||
- Streaming outputs
|
||||
- OpenAI-compatible API server
|
||||
- Support NVIDIA GPUs, AMD CPUs and GPUs, Intel CPUs and GPUs, PowerPC CPUs, TPU, and AWS Neuron
|
||||
@ -82,7 +77,6 @@ vLLM is flexible and easy to use with:
|
||||
- Multi-LoRA support
|
||||
|
||||
vLLM seamlessly supports most popular open-source models on HuggingFace, including:
|
||||
|
||||
- Transformer-like LLMs (e.g., Llama)
|
||||
- Mixture-of-Expert LLMs (e.g., Mixtral, Deepseek-V2 and V3)
|
||||
- Embedding Models (e.g., E5-Mistral)
|
||||
@ -99,7 +93,6 @@ pip install vllm
|
||||
```
|
||||
|
||||
Visit our [documentation](https://docs.vllm.ai/en/latest/) to learn more.
|
||||
|
||||
- [Installation](https://docs.vllm.ai/en/latest/getting_started/installation.html)
|
||||
- [Quickstart](https://docs.vllm.ai/en/latest/getting_started/quickstart.html)
|
||||
- [List of Supported Models](https://docs.vllm.ai/en/latest/models/supported_models.html)
|
||||
@ -116,7 +109,6 @@ vLLM is a community project. Our compute resources for development and testing a
|
||||
<!-- Note: Please sort them in alphabetical order. -->
|
||||
<!-- Note: Please keep these consistent with docs/community/sponsors.md -->
|
||||
Cash Donations:
|
||||
|
||||
- a16z
|
||||
- Dropbox
|
||||
- Sequoia Capital
|
||||
@ -124,8 +116,6 @@ Cash Donations:
|
||||
- ZhenFund
|
||||
|
||||
Compute Resources:
|
||||
|
||||
- Alibaba Cloud
|
||||
- AMD
|
||||
- Anyscale
|
||||
- AWS
|
||||
@ -165,7 +155,7 @@ If you use vLLM for your research, please cite our [paper](https://arxiv.org/abs
|
||||
## Contact Us
|
||||
|
||||
<!-- --8<-- [start:contact-us] -->
|
||||
- For technical questions and feature requests, please use GitHub [Issues](https://github.com/vllm-project/vllm/issues)
|
||||
- For technical questions and feature requests, please use GitHub [Issues](https://github.com/vllm-project/vllm/issues) or [Discussions](https://github.com/vllm-project/vllm/discussions)
|
||||
- For discussing with fellow users, please use the [vLLM Forum](https://discuss.vllm.ai)
|
||||
- For coordinating contributions and development, please use [Slack](https://slack.vllm.ai)
|
||||
- For security disclosures, please use GitHub's [Security Advisories](https://github.com/vllm-project/vllm/security/advisories) feature
|
||||
|
||||
36
RELEASE.md
36
RELEASE.md
@ -52,39 +52,3 @@ After branch cut, we approach finalizing the release branch with clear criteria
|
||||
* Release branch specific changes (e.g. change version identifiers or CI fixes)
|
||||
|
||||
Please note: **No feature work allowed for cherry picks**. All PRs that are considered for cherry-picks need to be merged on trunk, the only exception are Release branch specific changes.
|
||||
|
||||
## Manual validations
|
||||
|
||||
### E2E Performance Validation
|
||||
|
||||
Before each release, we perform end-to-end performance validation to ensure no regressions are introduced. This validation uses the [vllm-benchmark workflow](https://github.com/pytorch/pytorch-integration-testing/actions/workflows/vllm-benchmark.yml) on PyTorch CI.
|
||||
|
||||
**Current Coverage:**
|
||||
|
||||
* Models: Llama3, Llama4, and Mixtral
|
||||
* Hardware: NVIDIA H100 and AMD MI300x
|
||||
* _Note: Coverage may change based on new model releases and hardware availability_
|
||||
|
||||
**Performance Validation Process:**
|
||||
|
||||
**Step 1: Get Access**
|
||||
Request write access to the [pytorch/pytorch-integration-testing](https://github.com/pytorch/pytorch-integration-testing) repository to run the benchmark workflow.
|
||||
|
||||
**Step 2: Review Benchmark Setup**
|
||||
Familiarize yourself with the benchmark configurations:
|
||||
|
||||
* [CUDA setup](https://github.com/pytorch/pytorch-integration-testing/tree/main/vllm-benchmarks/benchmarks/cuda)
|
||||
* [ROCm setup](https://github.com/pytorch/pytorch-integration-testing/tree/main/vllm-benchmarks/benchmarks/rocm)
|
||||
|
||||
**Step 3: Run the Benchmark**
|
||||
Navigate to the [vllm-benchmark workflow](https://github.com/pytorch/pytorch-integration-testing/actions/workflows/vllm-benchmark.yml) and configure:
|
||||
|
||||
* **vLLM branch**: Set to the release branch (e.g., `releases/v0.9.2`)
|
||||
* **vLLM commit**: Set to the RC commit hash
|
||||
|
||||
**Step 4: Review Results**
|
||||
Once the workflow completes, benchmark results will be available on the [vLLM benchmark dashboard](https://hud.pytorch.org/benchmark/llms?repoName=vllm-project%2Fvllm) under the corresponding branch and commit.
|
||||
|
||||
**Step 5: Performance Comparison**
|
||||
Compare the current results against the previous release to verify no performance regressions have occurred. Here is an
|
||||
example of [v0.9.1 vs v0.9.2](https://hud.pytorch.org/benchmark/llms?startTime=Thu%2C%2017%20Apr%202025%2021%3A43%3A50%20GMT&stopTime=Wed%2C%2016%20Jul%202025%2021%3A43%3A50%20GMT&granularity=week&lBranch=releases/v0.9.1&lCommit=b6553be1bc75f046b00046a4ad7576364d03c835&rBranch=releases/v0.9.2&rCommit=a5dd03c1ebc5e4f56f3c9d3dc0436e9c582c978f&repoName=vllm-project%2Fvllm&benchmarkName=&modelName=All%20Models&backendName=All%20Backends&modeName=All%20Modes&dtypeName=All%20DType&deviceName=All%20Devices&archName=All%20Platforms).
|
||||
|
||||
45
SECURITY.md
45
SECURITY.md
@ -1,50 +1,13 @@
|
||||
# Security Policy
|
||||
|
||||
## Reporting security issues
|
||||
## Reporting a Vulnerability
|
||||
|
||||
Please report security issues privately using [the vulnerability submission form](https://github.com/vllm-project/vllm/security/advisories/new).
|
||||
If you believe you have found a security vulnerability in vLLM, we encourage you to let us know right away. We will investigate all legitimate reports and do our best to quickly fix the problem.
|
||||
|
||||
## Issue triage
|
||||
Please report security issues privately using [the vulnerability submission form](https://github.com/vllm-project/vllm/security/advisories/new). Reports will then be triaged by the [vulnerability management team](https://docs.vllm.ai/en/latest/contributing/vulnerability_management.html).
|
||||
|
||||
Reports will then be triaged by the [vulnerability management team](https://docs.vllm.ai/en/latest/contributing/vulnerability_management.html).
|
||||
|
||||
## Threat model
|
||||
---
|
||||
|
||||
Please see the [Security Guide in the vLLM documentation](https://docs.vllm.ai/en/latest/usage/security.html) for more information on vLLM's security assumptions and recommendations.
|
||||
|
||||
Please see [PyTorch's Security Policy](https://github.com/pytorch/pytorch/blob/main/SECURITY.md) for more information and recommendations on how to securely interact with models.
|
||||
|
||||
## Issue severity
|
||||
|
||||
We will determine the risk of each issue, taking into account our experience dealing with past issues, versions affected, common defaults, and use cases. We use the following severity categories:
|
||||
|
||||
### CRITICAL Severity
|
||||
|
||||
Vulnerabilities that allow remote attackers to execute arbitrary code, take full control of the system, or significantly compromise confidentiality, integrity, or availability without any interaction or privileges needed, examples include remote code execution via network, deserialization issues that allow exploit chains. Generally those issues which are rated as CVSS ≥ 9.0.
|
||||
|
||||
### HIGH Severity
|
||||
|
||||
Serious security flaws that allow elevated impact—like RCE in specific, limited contexts or significant data loss—but require advanced conditions or some trust, examples include RCE in advanced deployment modes (e.g. multi-node), or high impact issues where some sort of privileged network access is required. These issues typically have CVSS scores between 7.0 and 8.9
|
||||
|
||||
### MODERATE Severity
|
||||
|
||||
Vulnerabilities that cause denial of service or partial disruption, but do not allow arbitrary code execution or data breach and have limited impact. These issues have a CVSS rating between 4.0 and 6.9
|
||||
|
||||
### LOW Severity
|
||||
|
||||
Minor issues such as informational disclosures, logging errors, non-exploitable flaws, or weaknesses that require local or high-privilege access and offer negligible impact. Examples include side channel attacks or hash collisions. These issues often have CVSS scores less than 4.0
|
||||
|
||||
## Prenotification policy
|
||||
|
||||
For certain security issues of CRITICAL, HIGH, or MODERATE severity level, we may prenotify certain organizations or vendors that ship vLLM. The purpose of this prenotification is to allow for a coordinated release of fixes for severe issues.
|
||||
|
||||
* This prenotification will be in the form of a private email notification. It may also include adding security contacts to the GitHub security advisory, typically a few days before release.
|
||||
|
||||
* If you wish to be added to the prenotification group, please send an email copying all the members of the [vulnerability management team](https://docs.vllm.ai/en/latest/contributing/vulnerability_management.html). Each vendor contact will be analyzed on a case-by-case basis.
|
||||
|
||||
* Organizations and vendors who either ship or use vLLM, are eligible to join the prenotification group if they meet at least one of the following qualifications
|
||||
* Substantial internal deployment leveraging the upstream vLLM project.
|
||||
* Established internal security teams and comprehensive compliance measures.
|
||||
* Active and consistent contributions to the upstream vLLM project.
|
||||
|
||||
* We may withdraw organizations from receiving future prenotifications if they release fixes or any other information about issues before they are public. Group membership may also change based on policy refinements for who may be included.
|
||||
|
||||
@ -4,7 +4,7 @@ This README guides you through running benchmark tests with the extensive
|
||||
datasets supported on vLLM. It’s a living document, updated as new features and datasets
|
||||
become available.
|
||||
|
||||
## Dataset Overview
|
||||
**Dataset Overview**
|
||||
|
||||
<table style="width:100%; border-collapse: collapse;">
|
||||
<thead>
|
||||
@ -22,25 +22,6 @@ become available.
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>ShareGPT4V (Image)</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td>
|
||||
<code>wget https://huggingface.co/datasets/Lin-Chen/ShareGPT4V/blob/main/sharegpt4v_instruct_gpt4-vision_cap100k.json</code>
|
||||
<br>
|
||||
<div>Note that the images need to be downloaded separately. For example, to download COCO's 2017 Train images:</div>
|
||||
<code>wget http://images.cocodataset.org/zips/train2017.zip</code>
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>ShareGPT4Video (Video)</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td>
|
||||
<code>git clone https://huggingface.co/datasets/ShareGPT4Video/ShareGPT4Video</code>
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>BurstGPT</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
@ -48,7 +29,7 @@ become available.
|
||||
<td><code>wget https://github.com/HPMLL/BurstGPT/releases/download/v1.1/BurstGPT_without_fails_2.csv</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>Sonnet (deprecated)</strong></td>
|
||||
<td><strong>Sonnet</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td>Local file: <code>benchmarks/sonnet.txt</code></td>
|
||||
@ -59,18 +40,6 @@ become available.
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>synthetic</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>RandomMultiModal (Image/Video)</strong></td>
|
||||
<td style="text-align: center;">🟡</td>
|
||||
<td style="text-align: center;">🚧</td>
|
||||
<td><code>synthetic</code> </td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>Prefix Repetition</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>synthetic</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>HuggingFace-VisionArena</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
@ -110,24 +79,18 @@ become available.
|
||||
|
||||
🚧: to be supported
|
||||
|
||||
**Note**: HuggingFace dataset's `dataset-name` should be set to `hf`.
|
||||
For local `dataset-path`, please set `hf-name` to its Hugging Face ID like
|
||||
|
||||
```bash
|
||||
--dataset-path /datasets/VisionArena-Chat/ --hf-name lmarena-ai/VisionArena-Chat
|
||||
```
|
||||
|
||||
## 🚀 Example - Online Benchmark
|
||||
**Note**: HuggingFace dataset's `dataset-name` should be set to `hf`
|
||||
|
||||
---
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>🚀 Example - Online Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
First start serving your model
|
||||
|
||||
```bash
|
||||
vllm serve NousResearch/Hermes-3-Llama-3.1-8B
|
||||
vllm serve NousResearch/Hermes-3-Llama-3.1-8B --disable-log-requests
|
||||
```
|
||||
|
||||
Then run the benchmarking script
|
||||
@ -135,7 +98,7 @@ Then run the benchmarking script
|
||||
```bash
|
||||
# download dataset
|
||||
# wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--endpoint /v1/completions \
|
||||
@ -146,48 +109,48 @@ vllm bench serve \
|
||||
|
||||
If successful, you will see the following output
|
||||
|
||||
```text
|
||||
```
|
||||
============ Serving Benchmark Result ============
|
||||
Successful requests: 10
|
||||
Benchmark duration (s): 5.78
|
||||
Total input tokens: 1369
|
||||
Total generated tokens: 2212
|
||||
Request throughput (req/s): 1.73
|
||||
Output token throughput (tok/s): 382.89
|
||||
Total Token throughput (tok/s): 619.85
|
||||
Successful requests: 10
|
||||
Benchmark duration (s): 5.78
|
||||
Total input tokens: 1369
|
||||
Total generated tokens: 2212
|
||||
Request throughput (req/s): 1.73
|
||||
Output token throughput (tok/s): 382.89
|
||||
Total Token throughput (tok/s): 619.85
|
||||
---------------Time to First Token----------------
|
||||
Mean TTFT (ms): 71.54
|
||||
Median TTFT (ms): 73.88
|
||||
P99 TTFT (ms): 79.49
|
||||
Mean TTFT (ms): 71.54
|
||||
Median TTFT (ms): 73.88
|
||||
P99 TTFT (ms): 79.49
|
||||
-----Time per Output Token (excl. 1st token)------
|
||||
Mean TPOT (ms): 7.91
|
||||
Median TPOT (ms): 7.96
|
||||
P99 TPOT (ms): 8.03
|
||||
Mean TPOT (ms): 7.91
|
||||
Median TPOT (ms): 7.96
|
||||
P99 TPOT (ms): 8.03
|
||||
---------------Inter-token Latency----------------
|
||||
Mean ITL (ms): 7.74
|
||||
Median ITL (ms): 7.70
|
||||
P99 ITL (ms): 8.39
|
||||
Mean ITL (ms): 7.74
|
||||
Median ITL (ms): 7.70
|
||||
P99 ITL (ms): 8.39
|
||||
==================================================
|
||||
```
|
||||
|
||||
### Custom Dataset
|
||||
**Custom Dataset**
|
||||
|
||||
If the dataset you want to benchmark is not supported yet in vLLM, even then you can benchmark on it using `CustomDataset`. Your data needs to be in `.jsonl` format and needs to have "prompt" field per entry, e.g., data.jsonl
|
||||
|
||||
```json
|
||||
```
|
||||
{"prompt": "What is the capital of India?"}
|
||||
{"prompt": "What is the capital of Iran?"}
|
||||
{"prompt": "What is the capital of China?"}
|
||||
```
|
||||
```
|
||||
|
||||
```bash
|
||||
# start server
|
||||
VLLM_USE_V1=1 vllm serve meta-llama/Llama-3.1-8B-Instruct
|
||||
VLLM_USE_V1=1 vllm serve meta-llama/Llama-3.1-8B-Instruct --disable-log-requests
|
||||
```
|
||||
|
||||
```bash
|
||||
# run benchmarking script
|
||||
vllm bench serve --port 9001 --save-result --save-detailed \
|
||||
python3 benchmarks/benchmark_serving.py --port 9001 --save-result --save-detailed \
|
||||
--backend vllm \
|
||||
--model meta-llama/Llama-3.1-8B-Instruct \
|
||||
--endpoint /v1/completions \
|
||||
@ -203,17 +166,16 @@ vllm bench serve --port 9001 --save-result --save-detailed \
|
||||
|
||||
You can skip applying chat template if your data already has it by using `--custom-skip-chat-template`.
|
||||
|
||||
### VisionArena Benchmark for Vision Language Models
|
||||
**VisionArena Benchmark for Vision Language Models**
|
||||
|
||||
```bash
|
||||
# need a model with vision capability here
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct --disable-log-requests
|
||||
```
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--endpoint-type openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
@ -222,7 +184,7 @@ vllm bench serve \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### InstructCoder Benchmark with Speculative Decoding
|
||||
**InstructCoder Benchmark with Speculative Decoding**
|
||||
|
||||
``` bash
|
||||
VLLM_USE_V1=1 vllm serve meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
@ -232,25 +194,24 @@ VLLM_USE_V1=1 vllm serve meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
```
|
||||
|
||||
``` bash
|
||||
vllm bench serve \
|
||||
python3 benchmarks/benchmark_serving.py \
|
||||
--model meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
--dataset-name hf \
|
||||
--dataset-path likaixin/InstructCoder \
|
||||
--num-prompts 2048
|
||||
```
|
||||
|
||||
### Other HuggingFaceDataset Examples
|
||||
**Other HuggingFaceDataset Examples**
|
||||
|
||||
```bash
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct --disable-log-requests
|
||||
```
|
||||
|
||||
`lmms-lab/LLaVA-OneVision-Data`:
|
||||
**`lmms-lab/LLaVA-OneVision-Data`**
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--endpoint-type openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
@ -260,12 +221,11 @@ vllm bench serve \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
`Aeala/ShareGPT_Vicuna_unfiltered`:
|
||||
**`Aeala/ShareGPT_Vicuna_unfiltered`**
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--endpoint-type openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
@ -274,10 +234,10 @@ vllm bench serve \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
`AI-MO/aimo-validation-aime`:
|
||||
**`AI-MO/aimo-validation-aime`**
|
||||
|
||||
``` bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--model Qwen/QwQ-32B \
|
||||
--dataset-name hf \
|
||||
--dataset-path AI-MO/aimo-validation-aime \
|
||||
@ -285,23 +245,23 @@ vllm bench serve \
|
||||
--seed 42
|
||||
```
|
||||
|
||||
`philschmid/mt-bench`:
|
||||
**`philschmid/mt-bench`**
|
||||
|
||||
``` bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--model Qwen/QwQ-32B \
|
||||
--dataset-name hf \
|
||||
--dataset-path philschmid/mt-bench \
|
||||
--num-prompts 80
|
||||
```
|
||||
|
||||
### Running With Sampling Parameters
|
||||
**Running With Sampling Parameters**
|
||||
|
||||
When using OpenAI-compatible backends such as `vllm`, optional sampling
|
||||
parameters can be specified. Example client command:
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--endpoint /v1/completions \
|
||||
@ -313,34 +273,30 @@ vllm bench serve \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
### Running With Ramp-Up Request Rate
|
||||
**Running With Ramp-Up Request Rate**
|
||||
|
||||
The benchmark tool also supports ramping up the request rate over the
|
||||
duration of the benchmark run. This can be useful for stress testing the
|
||||
server or finding the maximum throughput that it can handle, given some latency budget.
|
||||
|
||||
Two ramp-up strategies are supported:
|
||||
|
||||
- `linear`: Increases the request rate linearly from a start value to an end value.
|
||||
- `exponential`: Increases the request rate exponentially.
|
||||
|
||||
The following arguments can be used to control the ramp-up:
|
||||
|
||||
- `--ramp-up-strategy`: The ramp-up strategy to use (`linear` or `exponential`).
|
||||
- `--ramp-up-start-rps`: The request rate at the beginning of the benchmark.
|
||||
- `--ramp-up-end-rps`: The request rate at the end of the benchmark.
|
||||
|
||||
</details>
|
||||
|
||||
## 📈 Example - Offline Throughput Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>📈 Example - Offline Throughput Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
```bash
|
||||
vllm bench throughput \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--dataset-name sonnet \
|
||||
--dataset-path vllm/benchmarks/sonnet.txt \
|
||||
@ -349,16 +305,16 @@ vllm bench throughput \
|
||||
|
||||
If successful, you will see the following output
|
||||
|
||||
```text
|
||||
```
|
||||
Throughput: 7.15 requests/s, 4656.00 total tokens/s, 1072.15 output tokens/s
|
||||
Total num prompt tokens: 5014
|
||||
Total num output tokens: 1500
|
||||
```
|
||||
|
||||
### VisionArena Benchmark for Vision Language Models
|
||||
**VisionArena Benchmark for Vision Language Models**
|
||||
|
||||
```bash
|
||||
vllm bench throughput \
|
||||
``` bash
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
@ -369,18 +325,18 @@ vllm bench throughput \
|
||||
|
||||
The `num prompt tokens` now includes image token counts
|
||||
|
||||
```text
|
||||
```
|
||||
Throughput: 2.55 requests/s, 4036.92 total tokens/s, 326.90 output tokens/s
|
||||
Total num prompt tokens: 14527
|
||||
Total num output tokens: 1280
|
||||
```
|
||||
|
||||
### InstructCoder Benchmark with Speculative Decoding
|
||||
**InstructCoder Benchmark with Speculative Decoding**
|
||||
|
||||
``` bash
|
||||
VLLM_WORKER_MULTIPROC_METHOD=spawn \
|
||||
VLLM_USE_V1=1 \
|
||||
vllm bench throughput \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--dataset-name=hf \
|
||||
--dataset-path=likaixin/InstructCoder \
|
||||
--model=meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
@ -393,18 +349,18 @@ vllm bench throughput \
|
||||
"prompt_lookup_min": 2}'
|
||||
```
|
||||
|
||||
```text
|
||||
```
|
||||
Throughput: 104.77 requests/s, 23836.22 total tokens/s, 10477.10 output tokens/s
|
||||
Total num prompt tokens: 261136
|
||||
Total num output tokens: 204800
|
||||
```
|
||||
|
||||
### Other HuggingFaceDataset Examples
|
||||
**Other HuggingFaceDataset Examples**
|
||||
|
||||
`lmms-lab/LLaVA-OneVision-Data`:
|
||||
**`lmms-lab/LLaVA-OneVision-Data`**
|
||||
|
||||
```bash
|
||||
vllm bench throughput \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
@ -414,10 +370,10 @@ vllm bench throughput \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
`Aeala/ShareGPT_Vicuna_unfiltered`:
|
||||
**`Aeala/ShareGPT_Vicuna_unfiltered`**
|
||||
|
||||
```bash
|
||||
vllm bench throughput \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
@ -426,10 +382,10 @@ vllm bench throughput \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
`AI-MO/aimo-validation-aime`:
|
||||
**`AI-MO/aimo-validation-aime`**
|
||||
|
||||
```bash
|
||||
vllm bench throughput \
|
||||
python3 benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/QwQ-32B \
|
||||
--backend vllm \
|
||||
--dataset-name hf \
|
||||
@ -438,12 +394,12 @@ vllm bench throughput \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
Benchmark with LoRA adapters:
|
||||
**Benchmark with LoRA Adapters**
|
||||
|
||||
``` bash
|
||||
# download dataset
|
||||
# wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
vllm bench throughput \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model meta-llama/Llama-2-7b-hf \
|
||||
--backend vllm \
|
||||
--dataset_path <your data path>/ShareGPT_V3_unfiltered_cleaned_split.json \
|
||||
@ -457,22 +413,20 @@ vllm bench throughput \
|
||||
|
||||
</details>
|
||||
|
||||
## 🛠️ Example - Structured Output Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>🛠️ Example - Structured Output Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
Benchmark the performance of structured output generation (JSON, grammar, regex).
|
||||
|
||||
### Server Setup
|
||||
**Server Setup**
|
||||
|
||||
```bash
|
||||
vllm serve NousResearch/Hermes-3-Llama-3.1-8B
|
||||
vllm serve NousResearch/Hermes-3-Llama-3.1-8B --disable-log-requests
|
||||
```
|
||||
|
||||
### JSON Schema Benchmark
|
||||
**JSON Schema Benchmark**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -484,7 +438,7 @@ python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### Grammar-based Generation Benchmark
|
||||
**Grammar-based Generation Benchmark**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -496,7 +450,7 @@ python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### Regex-based Generation Benchmark
|
||||
**Regex-based Generation Benchmark**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -507,7 +461,7 @@ python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### Choice-based Generation Benchmark
|
||||
**Choice-based Generation Benchmark**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -518,7 +472,7 @@ python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### XGrammar Benchmark Dataset
|
||||
**XGrammar Benchmark Dataset**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -531,16 +485,14 @@ python3 benchmarks/benchmark_serving_structured_output.py \
|
||||
|
||||
</details>
|
||||
|
||||
## 📚 Example - Long Document QA Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>📚 Example - Long Document QA Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
Benchmark the performance of long document question-answering with prefix caching.
|
||||
|
||||
### Basic Long Document QA Test
|
||||
**Basic Long Document QA Test**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_long_document_qa_throughput.py \
|
||||
@ -552,7 +504,7 @@ python3 benchmarks/benchmark_long_document_qa_throughput.py \
|
||||
--repeat-count 5
|
||||
```
|
||||
|
||||
### Different Repeat Modes
|
||||
**Different Repeat Modes**
|
||||
|
||||
```bash
|
||||
# Random mode (default) - shuffle prompts randomly
|
||||
@ -585,16 +537,14 @@ python3 benchmarks/benchmark_long_document_qa_throughput.py \
|
||||
|
||||
</details>
|
||||
|
||||
## 🗂️ Example - Prefix Caching Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>🗂️ Example - Prefix Caching Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
Benchmark the efficiency of automatic prefix caching.
|
||||
|
||||
### Fixed Prompt with Prefix Caching
|
||||
**Fixed Prompt with Prefix Caching**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_prefix_caching.py \
|
||||
@ -605,7 +555,7 @@ python3 benchmarks/benchmark_prefix_caching.py \
|
||||
--input-length-range 128:256
|
||||
```
|
||||
|
||||
### ShareGPT Dataset with Prefix Caching
|
||||
**ShareGPT Dataset with Prefix Caching**
|
||||
|
||||
```bash
|
||||
# download dataset
|
||||
@ -620,32 +570,16 @@ python3 benchmarks/benchmark_prefix_caching.py \
|
||||
--input-length-range 128:256
|
||||
```
|
||||
|
||||
### Prefix Repetition Dataset
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
--backend openai \
|
||||
--model meta-llama/Llama-2-7b-chat-hf \
|
||||
--dataset-name prefix_repetition \
|
||||
--num-prompts 100 \
|
||||
--prefix-repetition-prefix-len 512 \
|
||||
--prefix-repetition-suffix-len 128 \
|
||||
--prefix-repetition-num-prefixes 5 \
|
||||
--prefix-repetition-output-len 128
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
## ⚡ Example - Request Prioritization Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
<summary><b>⚡ Example - Request Prioritization Benchmark</b></summary>
|
||||
|
||||
<br/>
|
||||
|
||||
Benchmark the performance of request prioritization in vLLM.
|
||||
|
||||
### Basic Prioritization Test
|
||||
**Basic Prioritization Test**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_prioritization.py \
|
||||
@ -656,7 +590,7 @@ python3 benchmarks/benchmark_prioritization.py \
|
||||
--scheduling-policy priority
|
||||
```
|
||||
|
||||
### Multiple Sequences per Prompt
|
||||
**Multiple Sequences per Prompt**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_prioritization.py \
|
||||
@ -669,139 +603,3 @@ python3 benchmarks/benchmark_prioritization.py \
|
||||
```
|
||||
|
||||
</details>
|
||||
|
||||
## 👁️ Example - Multi-Modal Benchmark
|
||||
|
||||
<details>
|
||||
<summary>Show more</summary>
|
||||
|
||||
<br/>
|
||||
|
||||
Benchmark the performance of multi-modal requests in vLLM.
|
||||
|
||||
### Images (ShareGPT4V)
|
||||
|
||||
Start vLLM:
|
||||
|
||||
```bash
|
||||
python -m vllm.entrypoints.openai.api_server \
|
||||
--model Qwen/Qwen2.5-VL-7B-Instruct \
|
||||
--dtype bfloat16 \
|
||||
--limit-mm-per-prompt '{"image": 1}' \
|
||||
--allowed-local-media-path /path/to/sharegpt4v/images
|
||||
```
|
||||
|
||||
Send requests with images:
|
||||
|
||||
```bash
|
||||
python benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2.5-VL-7B-Instruct \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path /path/to/ShareGPT4V/sharegpt4v_instruct_gpt4-vision_cap100k.json \
|
||||
--num-prompts 100 \
|
||||
--save-result \
|
||||
--result-dir ~/vllm_benchmark_results \
|
||||
--save-detailed \
|
||||
--endpoint /v1/chat/completion
|
||||
```
|
||||
|
||||
### Videos (ShareGPT4Video)
|
||||
|
||||
Start vLLM:
|
||||
|
||||
```bash
|
||||
python -m vllm.entrypoints.openai.api_server \
|
||||
--model Qwen/Qwen2.5-VL-7B-Instruct \
|
||||
--dtype bfloat16 \
|
||||
--limit-mm-per-prompt '{"video": 1}' \
|
||||
--allowed-local-media-path /path/to/sharegpt4video/videos
|
||||
```
|
||||
|
||||
Send requests with videos:
|
||||
|
||||
```bash
|
||||
python benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2.5-VL-7B-Instruct \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path /path/to/ShareGPT4Video/llava_v1_5_mix665k_with_video_chatgpt72k_share4video28k.json \
|
||||
--num-prompts 100 \
|
||||
--save-result \
|
||||
--result-dir ~/vllm_benchmark_results \
|
||||
--save-detailed \
|
||||
--endpoint /v1/chat/completion
|
||||
```
|
||||
|
||||
### Synthetic Random Images (random-mm)
|
||||
|
||||
Generate synthetic image inputs alongside random text prompts to stress-test vision models without external datasets.
|
||||
|
||||
Notes:
|
||||
|
||||
- Works only with online benchmark via the OpenAI backend (`--backend openai-chat`) and endpoint `/v1/chat/completions`.
|
||||
- Video sampling is not yet implemented.
|
||||
|
||||
Start the server (example):
|
||||
|
||||
```bash
|
||||
vllm serve Qwen/Qwen2.5-VL-3B-Instruct \
|
||||
--dtype bfloat16 \
|
||||
--max-model-len 16384 \
|
||||
--limit-mm-per-prompt '{"image": 3, "video": 0}' \
|
||||
--mm-processor-kwargs max_pixels=1003520
|
||||
```
|
||||
|
||||
Benchmark. It is recommended to use the flag `--ignore-eos` to simulate real responses. You can set the size of the output via the arg `random-output-len`.
|
||||
|
||||
Ex.1: Fixed number of items and a single image resolution, enforcing generation of approx 40 tokens:
|
||||
|
||||
```bash
|
||||
vllm bench serve \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2.5-VL-3B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name random-mm \
|
||||
--num-prompts 100 \
|
||||
--max-concurrency 10 \
|
||||
--random-prefix-len 25 \
|
||||
--random-input-len 300 \
|
||||
--random-output-len 40 \
|
||||
--random-range-ratio 0.2 \
|
||||
--random-mm-base-items-per-request 2 \
|
||||
--random-mm-limit-mm-per-prompt '{"image": 3, "video": 0}' \
|
||||
--random-mm-bucket-config '{(224, 224, 1): 1.0}' \
|
||||
--request-rate inf \
|
||||
--ignore-eos \
|
||||
--seed 42
|
||||
```
|
||||
|
||||
The number of items per request can be controlled by passing multiple image buckets:
|
||||
|
||||
```bash
|
||||
--random-mm-base-items-per-request 2 \
|
||||
--random-mm-num-mm-items-range-ratio 0.5 \
|
||||
--random-mm-limit-mm-per-prompt '{"image": 4, "video": 0}' \
|
||||
--random-mm-bucket-config '{(256, 256, 1): 0.7, (720, 1280, 1): 0.3}' \
|
||||
```
|
||||
|
||||
Flags specific to `random-mm`:
|
||||
|
||||
- `--random-mm-base-items-per-request`: base number of multimodal items per request.
|
||||
- `--random-mm-num-mm-items-range-ratio`: vary item count uniformly in the closed integer range [floor(n·(1−r)), ceil(n·(1+r))]. Set r=0 to keep it fixed; r=1 allows 0 items.
|
||||
- `--random-mm-limit-mm-per-prompt`: per-modality hard caps, e.g. '{"image": 3, "video": 0}'.
|
||||
- `--random-mm-bucket-config`: dict mapping (H, W, T) → probability. Entries with probability 0 are removed; remaining probabilities are renormalized to sum to 1. Use T=1 for images. Set any T>1 for videos (video sampling not yet supported).
|
||||
|
||||
Behavioral notes:
|
||||
|
||||
- If the requested base item count cannot be satisfied under the provided per-prompt limits, the tool raises an error rather than silently clamping.
|
||||
|
||||
How sampling works:
|
||||
|
||||
- Determine per-request item count k by sampling uniformly from the integer range defined by `--random-mm-base-items-per-request` and `--random-mm-num-mm-items-range-ratio`, then clamp k to at most the sum of per-modality limits.
|
||||
- For each of the k items, sample a bucket (H, W, T) according to the normalized probabilities in `--random-mm-bucket-config`, while tracking how many items of each modality have been added.
|
||||
- If a modality (e.g., image) reaches its limit from `--random-mm-limit-mm-per-prompt`, all buckets of that modality are excluded and the remaining bucket probabilities are renormalized before continuing.
|
||||
This should be seen as an edge case, and if this behavior can be avoided by setting `--random-mm-limit-mm-per-prompt` to a large number. Note that this might result in errors due to engine config `--limit-mm-per-prompt`.
|
||||
- The resulting request contains synthetic image data in `multi_modal_data` (OpenAI Chat format). When `random-mm` is used with the OpenAI Chat backend, prompts remain text and MM content is attached via `multi_modal_data`.
|
||||
|
||||
</details>
|
||||
|
||||
@ -1,45 +1,56 @@
|
||||
#!/bin/bash
|
||||
|
||||
# This script aims to tune the best server parameter combinations to maximize throughput for given requirement.
|
||||
# See details in README (benchmarks/auto_tune/README.md).
|
||||
# This script aims to tune the best server parameter combinations to maximize throughput for given requirement.
|
||||
# The current server parameter combination is max_num_seqs and max_num_batched_tokens
|
||||
# It also supports additional requirement: e2e latency and prefix cache.
|
||||
|
||||
# Pre-requisite:
|
||||
# 1. Checkout to your branch, install/ update the correct running env. For TPU, activate conda env and install the corresponding torch, xla version.
|
||||
# 2. If the model is customized, replace the MODEL's config with the customized config.
|
||||
# 3. Set variables (ALL REQUIRED)
|
||||
# BASE: your directory for vllm repo
|
||||
# MODEL: the model served by vllm
|
||||
# SYSTEM: the hardware, choice TPU or GPU, for other systems, "get best profile" might not support.
|
||||
# TP: ways of tensor parallelism
|
||||
# DOWNLOAD_DIR: directory to download and load model weights.
|
||||
# INPUT_LEN: request input len
|
||||
# OUTPUT_LEN: request output len
|
||||
# MIN_CACHE_HIT_PCT: prefix cache rate
|
||||
# MAX_LATENCY_ALLOWED_MS: (e2e) latency requirement. If there's no latency requirement, set it to a large number like 1000000000
|
||||
# NUM_SEQS_LIST: a list of `max-num-seqs` you want to loop with.
|
||||
# NUM_BATCHED_TOKENS_LIST: a list of `max-num-batched-tokens` you want to loop with.
|
||||
# Note that the default NUM_SEQS_LIST and NUM_BATCHED_TOKENS_LIST are set for medium size input/output len, for extra short context (such as 20:20), you might need to include larger numbers in NUM_SEQS_LIST.
|
||||
# 4. Run the script, it might take a long time, you can use tmux to avoid the script stop if disconnection happens.
|
||||
# 5. The final result will be saved in RESULT file.
|
||||
|
||||
|
||||
# Example use cases
|
||||
# 1. Given input_len=1800, output_len=20, what's the best max_num_seqs and max_num_batched_tokens to get highest throughput?
|
||||
# Use INPUT_LEN=1800, OUTPUT_LEN=20, MIN_CACHE_HIT_PCT=0, MAX_LATENCY_ALLOWED_MS=100000000000
|
||||
# 2. If we have latency requirement to be lower than 500ms, what's the best server parameter?
|
||||
# Use INPUT_LEN=1800, OUTPUT_LEN=20, MIN_CACHE_HIT_PCT=0, MAX_LATENCY_ALLOWED_MS=500
|
||||
# 3. If we want to reach 60% prefix cache, what's the best server parameter?
|
||||
# Use INPUT_LEN=1800, OUTPUT_LEN=20, MIN_CACHE_HIT_PCT=60, MAX_LATENCY_ALLOWED_MS=500
|
||||
|
||||
TAG=$(date +"%Y_%m_%d_%H_%M")
|
||||
SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
|
||||
VLLM_LOGGING_LEVEL=${VLLM_LOGGING_LEVEL:-INFO}
|
||||
BASE=${BASE:-"$SCRIPT_DIR/../../.."}
|
||||
MODEL=${MODEL:-"meta-llama/Llama-3.1-8B-Instruct"}
|
||||
SYSTEM=${SYSTEM:-"TPU"}
|
||||
TP=${TP:-1}
|
||||
DOWNLOAD_DIR=${DOWNLOAD_DIR:-""}
|
||||
INPUT_LEN=${INPUT_LEN:-4000}
|
||||
OUTPUT_LEN=${OUTPUT_LEN:-16}
|
||||
MAX_MODEL_LEN=${MAX_MODEL_LEN:-4096}
|
||||
MIN_CACHE_HIT_PCT=${MIN_CACHE_HIT_PCT:-0}
|
||||
MAX_LATENCY_ALLOWED_MS=${MAX_LATENCY_ALLOWED_MS:-100000000000}
|
||||
NUM_SEQS_LIST=${NUM_SEQS_LIST:-"128 256"}
|
||||
NUM_BATCHED_TOKENS_LIST=${NUM_BATCHED_TOKENS_LIST:-"512 1024 2048 4096"}
|
||||
BASE=""
|
||||
MODEL="meta-llama/Llama-3.1-8B-Instruct"
|
||||
SYSTEM="TPU"
|
||||
TP=1
|
||||
DOWNLOAD_DIR=""
|
||||
INPUT_LEN=4000
|
||||
OUTPUT_LEN=16
|
||||
MIN_CACHE_HIT_PCT=0
|
||||
MAX_LATENCY_ALLOWED_MS=100000000000
|
||||
NUM_SEQS_LIST="128 256"
|
||||
NUM_BATCHED_TOKENS_LIST="512 1024 2048 4096"
|
||||
|
||||
LOG_FOLDER="$BASE/auto-benchmark/$TAG"
|
||||
RESULT="$LOG_FOLDER/result.txt"
|
||||
PROFILE_PATH="$LOG_FOLDER/profile"
|
||||
|
||||
echo "====================== AUTO TUNE PARAMETERS ===================="
|
||||
echo "SCRIPT_DIR=$SCRIPT_DIR"
|
||||
echo "BASE=$BASE"
|
||||
echo "MODEL=$MODEL"
|
||||
echo "SYSTEM=$SYSTEM"
|
||||
echo "TP=$TP"
|
||||
echo "DOWNLOAD_DIR=$DOWNLOAD_DIR"
|
||||
echo "INPUT_LEN=$INPUT_LEN"
|
||||
echo "OUTPUT_LEN=$OUTPUT_LEN"
|
||||
echo "MAX_MODEL_LEN=$MAX_MODEL_LEN"
|
||||
echo "MIN_CACHE_HIT_PCT=$MIN_CACHE_HIT_PCT"
|
||||
echo "MAX_LATENCY_ALLOWED_MS=$MAX_LATENCY_ALLOWED_MS"
|
||||
echo "NUM_SEQS_LIST=$NUM_SEQS_LIST"
|
||||
echo "NUM_BATCHED_TOKENS_LIST=$NUM_BATCHED_TOKENS_LIST"
|
||||
echo "VLLM_LOGGING_LEVEL=$VLLM_LOGGING_LEVEL"
|
||||
echo "RESULT_FILE=$RESULT"
|
||||
echo "====================== AUTO TUNEPARAMETERS ===================="
|
||||
echo "result file: $RESULT"
|
||||
echo "model: $MODEL"
|
||||
|
||||
rm -rf $LOG_FOLDER
|
||||
rm -rf $PROFILE_PATH
|
||||
@ -54,18 +65,10 @@ current_hash=$(git rev-parse HEAD)
|
||||
echo "hash:$current_hash" >> "$RESULT"
|
||||
echo "current_hash: $current_hash"
|
||||
|
||||
TOTAL_LEN=$((INPUT_LEN + OUTPUT_LEN))
|
||||
RED='\033[0;31m'
|
||||
if (( TOTAL_LEN > MAX_MODEL_LEN )); then
|
||||
echo -e "${RED}FAILED: INPUT_LEN($INPUT_LEN) + OUTPUT_LEN($OUTPUT_LEN) = $TOTAL_LEN, which is > MAX_MODEL_LEN = $MAX_MODEL_LEN.\033[0m" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
best_throughput=0
|
||||
best_max_num_seqs=0
|
||||
best_num_batched_tokens=0
|
||||
best_goodput=0
|
||||
best_request_rate=0
|
||||
|
||||
start_server() {
|
||||
local gpu_memory_utilization=$1
|
||||
@ -73,42 +76,26 @@ start_server() {
|
||||
local max_num_batched_tokens=$3
|
||||
local vllm_log=$4
|
||||
local profile_dir=$5
|
||||
|
||||
pkill -f vllm
|
||||
|
||||
pkill -if vllm
|
||||
|
||||
# Define the common arguments as a bash array.
|
||||
# Each argument and its value are separate elements.
|
||||
local common_args_array=(
|
||||
"$MODEL"
|
||||
"--disable-log-requests"
|
||||
"--port" "8004"
|
||||
"--gpu-memory-utilization" "$gpu_memory_utilization"
|
||||
"--max-num-seqs" "$max_num_seqs"
|
||||
"--max-num-batched-tokens" "$max_num_batched_tokens"
|
||||
"--tensor-parallel-size" "$TP"
|
||||
"--enable-prefix-caching"
|
||||
"--load-format" "dummy"
|
||||
"--download-dir" "$DOWNLOAD_DIR"
|
||||
"--max-model-len" "$MAX_MODEL_LEN"
|
||||
)
|
||||
|
||||
# Use the array expansion "${common_args_array[@]}"
|
||||
# This correctly passes each element as a separate argument.
|
||||
if [[ -n "$profile_dir" ]]; then
|
||||
# Start server with profiling enabled
|
||||
VLLM_USE_V1=1 VLLM_SERVER_DEV_MODE=1 VLLM_TORCH_PROFILER_DIR=$profile_dir \
|
||||
vllm serve "${common_args_array[@]}" > "$vllm_log" 2>&1 &
|
||||
else
|
||||
# Start server without profiling
|
||||
VLLM_USE_V1=1 VLLM_SERVER_DEV_MODE=1 \
|
||||
vllm serve "${common_args_array[@]}" > "$vllm_log" 2>&1 &
|
||||
fi
|
||||
VLLM_USE_V1=1 VLLM_SERVER_DEV_MODE=1 VLLM_TORCH_PROFILER_DIR=$profile_dir vllm serve $MODEL \
|
||||
--disable-log-requests \
|
||||
--port 8004 \
|
||||
--gpu-memory-utilization $gpu_memory_utilization \
|
||||
--max-num-seqs $max_num_seqs \
|
||||
--max-num-batched-tokens $max_num_batched_tokens \
|
||||
--tensor-parallel-size $TP \
|
||||
--enable-prefix-caching \
|
||||
--load-format dummy \
|
||||
--download-dir "$DOWNLOAD_DIR" \
|
||||
--max-model-len $(( INPUT_LEN+OUTPUT_LEN )) > "$vllm_log" 2>&1 &
|
||||
|
||||
# wait for 10 minutes...
|
||||
server_started=0
|
||||
for i in {1..60}; do
|
||||
for i in {1..60}; do
|
||||
RESPONSE=$(curl -s -X GET "http://0.0.0.0:8004/health" -w "%{http_code}" -o /dev/stdout)
|
||||
STATUS_CODE=$(echo "$RESPONSE" | tail -n 1)
|
||||
STATUS_CODE=$(echo "$RESPONSE" | tail -n 1)
|
||||
if [[ "$STATUS_CODE" -eq 200 ]]; then
|
||||
server_started=1
|
||||
break
|
||||
@ -116,7 +103,6 @@ start_server() {
|
||||
sleep 10
|
||||
fi
|
||||
done
|
||||
|
||||
if (( ! server_started )); then
|
||||
echo "server did not start within 10 minutes. Please check server log at $vllm_log".
|
||||
return 1
|
||||
@ -125,20 +111,37 @@ start_server() {
|
||||
fi
|
||||
}
|
||||
|
||||
update_best_profile() {
|
||||
local profile_dir=$1
|
||||
local profile_index=$2
|
||||
sorted_paths=($(find "$profile_dir" -maxdepth 1 -not -path "$profile_dir" | sort))
|
||||
selected_profile_file=
|
||||
if [[ "$SYSTEM" == "TPU" ]]; then
|
||||
selected_profile_file="${sorted_paths[$profile_index]}/*.xplane.pb"
|
||||
fi
|
||||
if [[ "$SYSTEM" == "GPU" ]]; then
|
||||
selected_profile_file="${sorted_paths[$profile_index]}"
|
||||
fi
|
||||
rm -f $PROFILE_PATH/*
|
||||
cp $selected_profile_file $PROFILE_PATH
|
||||
}
|
||||
|
||||
run_benchmark() {
|
||||
local max_num_seqs=$1
|
||||
local max_num_batched_tokens=$2
|
||||
local gpu_memory_utilization=$3
|
||||
echo "max_num_seq: $max_num_seqs, max_num_batched_tokens: $max_num_batched_tokens"
|
||||
local vllm_log="$LOG_FOLDER/vllm_log_${max_num_seqs}_${max_num_batched_tokens}.txt"
|
||||
local profile_dir="$LOG_FOLDER/profile_${max_num_seqs}_${max_num_batched_tokens}"
|
||||
echo "vllm_log: $vllm_log"
|
||||
echo
|
||||
rm -f $vllm_log
|
||||
pkill -if vllm
|
||||
mkdir -p $profile_dir
|
||||
pkill -f vllm
|
||||
local profile_index=0
|
||||
|
||||
echo "starting server..."
|
||||
# Call start_server without a profile_dir to avoid profiling overhead
|
||||
start_server $gpu_memory_utilization $max_num_seqs $max_num_batched_tokens $vllm_log ""
|
||||
start_server $gpu_memory_utilization $max_num_seqs $max_num_batched_tokens $vllm_log $profile_dir
|
||||
result=$?
|
||||
if [[ "$result" -eq 1 ]]; then
|
||||
echo "server failed to start. gpu_memory_utilization:$gpu_memory_utilization, max_num_seqs:$max_num_seqs, max_num_batched_tokens: $max_num_batched_tokens"
|
||||
@ -146,19 +149,17 @@ run_benchmark() {
|
||||
echo "server started."
|
||||
fi
|
||||
echo
|
||||
|
||||
|
||||
echo "run benchmark test..."
|
||||
meet_latency_requirement=0
|
||||
# get a basic qps by using request-rate inf
|
||||
bm_log="$LOG_FOLDER/bm_log_${max_num_seqs}_${max_num_batched_tokens}_requestrate_inf.txt"
|
||||
prefix_len=$(( INPUT_LEN * MIN_CACHE_HIT_PCT / 100 ))
|
||||
adjusted_input_len=$(( INPUT_LEN - prefix_len ))
|
||||
# --profile flag is removed from this call
|
||||
vllm bench serve \
|
||||
python benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $MODEL \
|
||||
--dataset-name random \
|
||||
--random-input-len $adjusted_input_len \
|
||||
--random-input-len $INPUT_LEN \
|
||||
--random-output-len $OUTPUT_LEN \
|
||||
--ignore-eos \
|
||||
--disable-tqdm \
|
||||
@ -167,7 +168,8 @@ run_benchmark() {
|
||||
--goodput e2el:$MAX_LATENCY_ALLOWED_MS \
|
||||
--num-prompts 1000 \
|
||||
--random-prefix-len $prefix_len \
|
||||
--port 8004 &> "$bm_log"
|
||||
--port 8004 \
|
||||
--profile &> "$bm_log"
|
||||
throughput=$(grep "Request throughput (req/s):" "$bm_log" | sed 's/[^0-9.]//g')
|
||||
e2el=$(grep "P99 E2EL (ms):" "$bm_log" | awk '{print $NF}')
|
||||
goodput=$(grep "Request goodput (req/s):" "$bm_log" | sed 's/[^0-9.]//g')
|
||||
@ -181,15 +183,16 @@ run_benchmark() {
|
||||
# start from request-rate as int(throughput) + 1
|
||||
request_rate=$((${throughput%.*} + 1))
|
||||
while ((request_rate > 0)); do
|
||||
profile_index=$((profile_index+1))
|
||||
# clear prefix cache
|
||||
curl -X POST http://0.0.0.0:8004/reset_prefix_cache
|
||||
sleep 5
|
||||
bm_log="$LOG_FOLDER/bm_log_${max_num_seqs}_${max_num_batched_tokens}_requestrate_${request_rate}.txt"
|
||||
vllm bench serve \
|
||||
python benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $MODEL \
|
||||
--dataset-name random \
|
||||
--random-input-len $adjusted_input_len \
|
||||
--random-input-len $INPUT_LEN \
|
||||
--random-output-len $OUTPUT_LEN \
|
||||
--ignore-eos \
|
||||
--disable-tqdm \
|
||||
@ -218,7 +221,12 @@ run_benchmark() {
|
||||
best_max_num_seqs=$max_num_seqs
|
||||
best_num_batched_tokens=$max_num_batched_tokens
|
||||
best_goodput=$goodput
|
||||
best_request_rate=$request_rate
|
||||
if [[ "$SYSTEM" == "TPU" ]]; then
|
||||
update_best_profile "$profile_dir/plugins/profile" $profile_index
|
||||
fi
|
||||
if [[ "$SYSTEM" == "GPU" ]]; then
|
||||
update_best_profile "$profile_dir" $profile_index
|
||||
fi
|
||||
fi
|
||||
else
|
||||
echo "max_num_seqs: $max_num_seqs, max_num_batched_tokens: $max_num_batched_tokens does not meet latency requirement ${MAX_LATENCY_ALLOWED_MS}"
|
||||
@ -227,9 +235,9 @@ run_benchmark() {
|
||||
|
||||
echo "best_max_num_seqs: $best_max_num_seqs, best_num_batched_tokens: $best_num_batched_tokens, best_throughput: $best_throughput"
|
||||
|
||||
pkill -if vllm
|
||||
pkill vllm
|
||||
sleep 10
|
||||
echo "===================="
|
||||
printf '=%.0s' $(seq 1 20)
|
||||
return 0
|
||||
}
|
||||
|
||||
@ -240,8 +248,7 @@ read -r -a num_batched_tokens_list <<< "$NUM_BATCHED_TOKENS_LIST"
|
||||
gpu_memory_utilization=0.98
|
||||
find_gpu_memory_utilization=0
|
||||
while (( $(echo "$gpu_memory_utilization >= 0.9" | bc -l) )); do
|
||||
# Pass empty string for profile_dir argument
|
||||
start_server $gpu_memory_utilization "${num_seqs_list[-1]}" "${num_batched_tokens_list[-1]}" "$LOG_FOLDER/vllm_log_gpu_memory_utilization_$gpu_memory_utilization.log" ""
|
||||
start_server $gpu_memory_utilization "${num_seqs_list[-1]}" "${num_batched_tokens_list[-1]}" "$LOG_FOLDER/vllm_log_gpu_memory_utilization_$gpu_memory_utilization.log"
|
||||
result=$?
|
||||
if [[ "$result" -eq 0 ]]; then
|
||||
find_gpu_memory_utilization=1
|
||||
@ -264,45 +271,6 @@ for num_seqs in "${num_seqs_list[@]}"; do
|
||||
done
|
||||
done
|
||||
echo "finish permutations"
|
||||
|
||||
# =================================================================================
|
||||
# FINAL PROFILING RUN FOR THE BEST CONFIGURATION
|
||||
# =================================================================================
|
||||
if (( $(echo "$best_throughput > 0" | bc -l) )); then
|
||||
echo
|
||||
echo "Benchmark tuning finished. Now running profiling on the best configuration found..."
|
||||
echo "Best config: max_num_seqs: $best_max_num_seqs, max_num_batched_tokens: $best_num_batched_tokens, throughput: $best_throughput"
|
||||
echo
|
||||
|
||||
vllm_log="$LOG_FOLDER/vllm_log_BEST_PROFILE.txt"
|
||||
bm_log="$LOG_FOLDER/bm_log_BEST_PROFILE.txt"
|
||||
|
||||
# Start server with the best params and profiling ENABLED
|
||||
echo "Starting server for profiling..."
|
||||
start_server $gpu_memory_utilization $best_max_num_seqs $best_num_batched_tokens "$vllm_log" "$PROFILE_PATH"
|
||||
|
||||
# Run benchmark with the best params and the --profile flag
|
||||
echo "Running benchmark with profiling..."
|
||||
prefix_len=$(( INPUT_LEN * MIN_CACHE_HIT_PCT / 100 ))
|
||||
adjusted_input_len=$(( INPUT_LEN - prefix_len ))
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--model $MODEL \
|
||||
--dataset-name random \
|
||||
--random-input-len $adjusted_input_len \
|
||||
--random-output-len $OUTPUT_LEN \
|
||||
--ignore-eos \
|
||||
--disable-tqdm \
|
||||
--request-rate $best_request_rate \
|
||||
--percentile-metrics ttft,tpot,itl,e2el \
|
||||
--goodput e2el:$MAX_LATENCY_ALLOWED_MS \
|
||||
--num-prompts 100 \
|
||||
--random-prefix-len $prefix_len \
|
||||
--port 8004 \
|
||||
--profile &> "$bm_log"
|
||||
else
|
||||
echo "No configuration met the latency requirements. Skipping final profiling run."
|
||||
fi
|
||||
pkill -if vllm
|
||||
echo "best_max_num_seqs: $best_max_num_seqs, best_num_batched_tokens: $best_num_batched_tokens, best_throughput: $best_throughput, profile saved in: $PROFILE_PATH"
|
||||
echo "best_max_num_seqs: $best_max_num_seqs, best_num_batched_tokens: $best_num_batched_tokens, best_throughput: $best_throughput, profile saved in: $PROFILE_PATH" >> "$RESULT"
|
||||
|
||||
@ -1,151 +0,0 @@
|
||||
# Automated vLLM Server Parameter Tuning
|
||||
|
||||
This script automates the process of finding the optimal server parameter combination (`max-num-seqs` and `max-num-batched-tokens`) to maximize throughput for a vLLM server. It also supports additional constraints such as E2E latency and prefix cache hit rate.
|
||||
|
||||
## Table of Contents
|
||||
|
||||
- [Prerequisites](#prerequisites)
|
||||
- [Configuration](#configuration)
|
||||
- [How to Run](#how-to-run)
|
||||
- [Example Use Cases](#example-use-cases)
|
||||
- [Output](#output)
|
||||
- [How It Works](#how-it-works)
|
||||
|
||||
## Prerequisites
|
||||
|
||||
Before running the script, please ensure the following steps are completed:
|
||||
|
||||
1. **Clone vLLM & Set Up Branch**: Clone the vLLM repository and check out to your desired branch.
|
||||
|
||||
```bash
|
||||
git clone https://github.com/vllm-project/vllm.git
|
||||
cd vllm
|
||||
# git checkout <your-branch>
|
||||
```
|
||||
|
||||
1. **Install Environment**: Install or update the correct running environment. For TPU usage, activate your `conda` environment and install the corresponding `torch` and `torch_xla` versions.
|
||||
|
||||
2. **Model Configuration**: If you are using a customized model, ensure its configuration files are correctly placed and accessible.
|
||||
|
||||
## Configuration
|
||||
|
||||
You must set the following variables at the top of the script before execution.
|
||||
|
||||
Note: You can also override the default values below via environment variables when running the script.
|
||||
|
||||
```bash
|
||||
MODEL=meta-llama/Llama-3.3-70B-Instruct SYSTEM=TPU TP=8 DOWNLOAD_DIR='' INPUT_LEN=128 OUTPUT_LEN=2048 MAX_MODEL_LEN=2300 MIN_CACHE_HIT_PCT=0 MAX_LATENCY_ALLOWED_MS=100000000000 NUM_SEQS_LIST="128 256" NUM_BATCHED_TOKENS_LIST="1024 2048 4096" VLLM_LOGGING_LEVEL=DEBUG bash auto_tune.sh
|
||||
```
|
||||
|
||||
| Variable | Description | Example Value |
|
||||
| --- | --- | --- |
|
||||
| `BASE` | **Required.** The absolute path to the parent directory of your vLLM repository directory. | `"$HOME"` |
|
||||
| `MODEL` | **Required.** The Hugging Face model identifier to be served by vllm. | `"meta-llama/Llama-3.1-8B-Instruct"` |
|
||||
| `SYSTEM`| **Required.** The hardware you are running on. Choices: `TPU` or `GPU`. (For other systems, it might not support saving profiles) | `"TPU"` |
|
||||
| `TP` | **Required.** The tensor-parallelism size. | `1` |
|
||||
| `DOWNLOAD_DIR` | **Required.** Directory to download and load model weights from. | `""` (default download path) |
|
||||
| `INPUT_LEN` | **Required.** Request input length. | `4000` |
|
||||
| `OUTPUT_LEN` | **Required.** Request output length. | `16` |
|
||||
| `MAX_MODEL_LEN` | **Required.** Max model length. | `4096` |
|
||||
| `MIN_CACHE_HIT_PCT` | Prefix cache hit rate in percentage (0-100). Set to `0` to disable. | `60` |
|
||||
| `MAX_LATENCY_ALLOWED_MS` | The maximum allowed P99 end-to-end latency in milliseconds. Set to a very large number (e.g., `100000000000`) to effectively ignore the latency constraint. | `500` |
|
||||
| `NUM_SEQS_LIST` | A space-separated string of `max-num-seqs` values to test. | `"128 256"` |
|
||||
| `NUM_BATCHED_TOKENS_LIST` | A space-separated string of `max-num-batched-tokens` values to test. | `"1024 2048 4096"` |
|
||||
|
||||
**Note**: The default `NUM_SEQS_LIST` and `NUM_BATCHED_TOKENS_LIST` are set for medium-sized inputs/outputs. For very short contexts (e.g., 20 input, 20 output tokens), you may need to test larger values for `max-num-seqs`.
|
||||
|
||||
## How to Run
|
||||
|
||||
1. **Configure**: Edit the script and set the variables in the [Configuration](#configuration) section.
|
||||
2. **Execute**: Run the script. Since the process can take a long time, it is highly recommended to use a terminal multiplexer like `tmux` or `screen` to prevent the script from stopping if your connection is lost.
|
||||
|
||||
```bash
|
||||
cd <FOLDER_OF_THIS_SCRIPT>
|
||||
bash auto_tune.sh
|
||||
```
|
||||
|
||||
Please note that the `bash auto_tune.sh` command cannot contain full or partial path with keyword `vllm`, otherwise `pkill -f vllm` command will also kill this script itself.
|
||||
|
||||
## Example Use Cases
|
||||
|
||||
Here are a few examples of how to configure the script for different goals:
|
||||
|
||||
### 1. Maximize Throughput (No Latency Constraint)
|
||||
|
||||
- **Goal**: Find the best `max-num-seqs` and `max-num-batched-tokens` to get the highest possible throughput for 1800 input tokens and 20 output tokens.
|
||||
- **Configuration**:
|
||||
|
||||
```bash
|
||||
INPUT_LEN=1800
|
||||
OUTPUT_LEN=20
|
||||
MAX_MODEL_LEN=2048
|
||||
MIN_CACHE_HIT_PCT=0
|
||||
MAX_LATENCY_ALLOWED_MS=100000000000 # A very large number
|
||||
```
|
||||
|
||||
#### 2. Maximize Throughput with a Latency Requirement
|
||||
|
||||
- **Goal**: Find the best server parameters when P99 end-to-end latency must be below 500ms.
|
||||
- **Configuration**:
|
||||
|
||||
```bash
|
||||
INPUT_LEN=1800
|
||||
OUTPUT_LEN=20
|
||||
MAX_MODEL_LEN=2048
|
||||
MIN_CACHE_HIT_PCT=0
|
||||
MAX_LATENCY_ALLOWED_MS=500
|
||||
```
|
||||
|
||||
#### 3. Maximize Throughput with Prefix Caching and Latency Requirements
|
||||
|
||||
- **Goal**: Find the best server parameters assuming a 60% prefix cache hit rate and a latency requirement of 500ms.
|
||||
- **Configuration**:
|
||||
|
||||
```bash
|
||||
INPUT_LEN=1800
|
||||
OUTPUT_LEN=20
|
||||
MAX_MODEL_LEN=2048
|
||||
MIN_CACHE_HIT_PCT=60
|
||||
MAX_LATENCY_ALLOWED_MS=500
|
||||
```
|
||||
|
||||
## Output
|
||||
|
||||
After the script finishes, you will find the results in a new, timestamped directory created inside `$BASE/auto-benchmark/`.
|
||||
|
||||
- **Log Files**: The directory (`$BASE/auto-benchmark/YYYY_MM_DD_HH_MM/`) contains detailed logs for each run:
|
||||
- `vllm_log_...txt`: The log output from the vLLM server for each parameter combination.
|
||||
- `bm_log_...txt`: The log output from the `vllm bench serve` command for each benchmark run.
|
||||
|
||||
- **Final Result Summary**: A file named `result.txt` is created in the log directory. It contains a summary of each tested combination and concludes with the overall best parameters found.
|
||||
|
||||
```text
|
||||
# Example result.txt content
|
||||
hash:a1b2c3d4...
|
||||
max_num_seqs: 128, max_num_batched_tokens: 2048, request_rate: 10.0, e2el: 450.5, throughput: 9.8, goodput: 9.8
|
||||
max_num_seqs: 128, max_num_batched_tokens: 4096 does not meet latency requirement 500
|
||||
...
|
||||
best_max_num_seqs: 256, best_num_batched_tokens: 2048, best_throughput: 12.5, profile saved in: /home/user/vllm/auto-benchmark/2024_08_01_10_30/profile
|
||||
```
|
||||
|
||||
If it cannot find the best parameters, the final row will be `best_max_num_seqs: 0, best_num_batched_tokens: 0, best_throughput: 0`. This can be due to either the server not starting properly, or the latency requirement being too strict.
|
||||
|
||||
- **Profiler Trace**: A directory named `profile` is created inside the log directory. It contains the profiler trace file (e.g., `.xplane.pb` for TPU or a `.json` trace for GPU) from the single best-performing run.
|
||||
|
||||
## How It Works
|
||||
|
||||
The script follows a systematic process to find the optimal parameters:
|
||||
|
||||
1. **Find Max GPU Memory Utilization**: The script first determines the highest safe `gpu-memory-utilization` (starting from 0.98 and decreasing) that does not cause an Out-Of-Memory (OOM) error when launching the server. This ensures the benchmark runs use the maximum available memory without crashing.
|
||||
|
||||
2. **Iterate and Benchmark**: It then enters a nested loop, iterating through every combination of `max-num-seqs` and `max-num-batched-tokens` provided in the configuration lists.
|
||||
|
||||
3. **Latency-Aware Throughput Search**: For each parameter combination:
|
||||
- The vLLM server is started.
|
||||
- A benchmark is first run with an infinite request rate (`--request-rate inf`).
|
||||
- If the resulting P99 E2E latency is within the `MAX_LATENCY_ALLOWED_MS` limit, this throughput is considered the maximum for this configuration.
|
||||
- If the latency is too high, the script performs a search by iteratively decreasing the request rate until the latency constraint is met. This finds the highest sustainable throughput for the given parameters and latency requirement.
|
||||
|
||||
4. **Track Best Result**: Throughout the process, the script tracks the parameter combination that has yielded the highest valid throughput so far.
|
||||
|
||||
5. **Profile Collection**: For the best-performing run, the script saves the vLLM profiler output, which can be used for deep-dive performance analysis with tools like TensorBoard.
|
||||
@ -31,10 +31,9 @@ class RequestFuncInput:
|
||||
model_name: Optional[str] = None
|
||||
logprobs: Optional[int] = None
|
||||
extra_body: Optional[dict] = None
|
||||
multi_modal_content: Optional[dict | list[dict]] = None
|
||||
multi_modal_content: Optional[dict] = None
|
||||
ignore_eos: bool = False
|
||||
language: Optional[str] = None
|
||||
request_id: Optional[str] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
@ -72,9 +71,6 @@ async def async_request_tgi(
|
||||
"inputs": request_func_input.prompt,
|
||||
"parameters": params,
|
||||
}
|
||||
headers = None
|
||||
if request_func_input.request_id:
|
||||
headers = {"x-request-id": request_func_input.request_id}
|
||||
output = RequestFuncOutput()
|
||||
output.prompt_len = request_func_input.prompt_len
|
||||
if request_func_input.ignore_eos:
|
||||
@ -86,9 +82,7 @@ async def async_request_tgi(
|
||||
st = time.perf_counter()
|
||||
most_recent_timestamp = st
|
||||
try:
|
||||
async with session.post(
|
||||
url=api_url, json=payload, headers=headers
|
||||
) as response:
|
||||
async with session.post(url=api_url, json=payload) as response:
|
||||
if response.status == 200:
|
||||
async for chunk_bytes in response.content:
|
||||
chunk_bytes = chunk_bytes.strip()
|
||||
@ -151,9 +145,6 @@ async def async_request_trt_llm(
|
||||
}
|
||||
if request_func_input.ignore_eos:
|
||||
payload["min_length"] = request_func_input.output_len
|
||||
headers = None
|
||||
if request_func_input.request_id:
|
||||
headers = {"x-request-id": request_func_input.request_id}
|
||||
output = RequestFuncOutput()
|
||||
output.prompt_len = request_func_input.prompt_len
|
||||
|
||||
@ -161,9 +152,7 @@ async def async_request_trt_llm(
|
||||
st = time.perf_counter()
|
||||
most_recent_timestamp = st
|
||||
try:
|
||||
async with session.post(
|
||||
url=api_url, json=payload, headers=headers
|
||||
) as response:
|
||||
async with session.post(url=api_url, json=payload) as response:
|
||||
if response.status == 200:
|
||||
async for chunk_bytes in response.content:
|
||||
chunk_bytes = chunk_bytes.strip()
|
||||
@ -222,8 +211,6 @@ async def async_request_deepspeed_mii(
|
||||
"top_p": 1.0,
|
||||
}
|
||||
headers = {"Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}"}
|
||||
if request_func_input.request_id:
|
||||
headers["x-request-id"] = request_func_input.request_id
|
||||
|
||||
output = RequestFuncOutput()
|
||||
output.prompt_len = request_func_input.prompt_len
|
||||
@ -296,8 +283,6 @@ async def async_request_openai_completions(
|
||||
if request_func_input.extra_body:
|
||||
payload.update(request_func_input.extra_body)
|
||||
headers = {"Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}"}
|
||||
if request_func_input.request_id:
|
||||
headers["x-request-id"] = request_func_input.request_id
|
||||
|
||||
output = RequestFuncOutput()
|
||||
output.prompt_len = request_func_input.prompt_len
|
||||
@ -379,15 +364,7 @@ async def async_request_openai_chat_completions(
|
||||
) as session:
|
||||
content = [{"type": "text", "text": request_func_input.prompt}]
|
||||
if request_func_input.multi_modal_content:
|
||||
mm_content = request_func_input.multi_modal_content
|
||||
if isinstance(mm_content, list):
|
||||
content.extend(mm_content)
|
||||
elif isinstance(mm_content, dict):
|
||||
content.append(mm_content)
|
||||
else:
|
||||
raise TypeError(
|
||||
"multi_modal_content must be a dict or list[dict] for openai-chat"
|
||||
)
|
||||
content.append(request_func_input.multi_modal_content)
|
||||
payload = {
|
||||
"model": request_func_input.model_name
|
||||
if request_func_input.model_name
|
||||
@ -410,8 +387,6 @@ async def async_request_openai_chat_completions(
|
||||
"Content-Type": "application/json",
|
||||
"Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}",
|
||||
}
|
||||
if request_func_input.request_id:
|
||||
headers["x-request-id"] = request_func_input.request_id
|
||||
|
||||
output = RequestFuncOutput()
|
||||
output.prompt_len = request_func_input.prompt_len
|
||||
@ -508,8 +483,6 @@ async def async_request_openai_audio(
|
||||
headers = {
|
||||
"Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}",
|
||||
}
|
||||
if request_func_input.request_id:
|
||||
headers["x-request-id"] = request_func_input.request_id
|
||||
|
||||
# Send audio file
|
||||
def to_bytes(y, sr):
|
||||
@ -518,10 +491,7 @@ async def async_request_openai_audio(
|
||||
buffer.seek(0)
|
||||
return buffer
|
||||
|
||||
mm_audio = request_func_input.multi_modal_content
|
||||
if not isinstance(mm_audio, dict) or "audio" not in mm_audio:
|
||||
raise TypeError("multi_modal_content must be a dict containing 'audio'")
|
||||
with to_bytes(*mm_audio["audio"]) as f:
|
||||
with to_bytes(*request_func_input.multi_modal_content["audio"]) as f:
|
||||
form = aiohttp.FormData()
|
||||
form.add_field("file", f, content_type="audio/wav")
|
||||
for key, value in payload.items():
|
||||
|
||||
@ -1,74 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import gc
|
||||
|
||||
from tabulate import tabulate
|
||||
|
||||
from benchmark_utils import TimeCollector
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
from vllm.v1.core.block_pool import BlockPool
|
||||
|
||||
|
||||
def main(args):
|
||||
rows = []
|
||||
for allocate_block in args.allocate_blocks:
|
||||
# Enforce a GC collect ahead to minimize the impact among runs
|
||||
gc.collect()
|
||||
block_pool = BlockPool(num_gpu_blocks=args.num_gpu_blocks, enable_caching=True)
|
||||
|
||||
get_blocks_times = TimeCollector(TimeCollector.US)
|
||||
free_blocks_times = TimeCollector(TimeCollector.US)
|
||||
for _ in range(args.num_iteration):
|
||||
with get_blocks_times:
|
||||
blocks = block_pool.get_new_blocks(allocate_block)
|
||||
with free_blocks_times:
|
||||
block_pool.free_blocks(blocks)
|
||||
|
||||
rows.append(
|
||||
[get_blocks_times.cnt, args.num_gpu_blocks, allocate_block]
|
||||
+ get_blocks_times.dump_avg_max()
|
||||
+ free_blocks_times.dump_avg_max()
|
||||
)
|
||||
|
||||
print(
|
||||
tabulate(
|
||||
rows,
|
||||
headers=[
|
||||
"Iterations",
|
||||
"Total\nBlocks",
|
||||
"Allocated\nBlocks",
|
||||
"Get Blocks\nAvg (us)",
|
||||
"Get Blocks\nMax (us)",
|
||||
"Free Blocks\nAvg (us)",
|
||||
"Free Blocks\nMax (us)",
|
||||
],
|
||||
tablefmt="grid",
|
||||
floatfmt=".3f",
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def invoke_main() -> None:
|
||||
parser = FlexibleArgumentParser(
|
||||
description="Benchmark the performance of BlockPool for KV Cache."
|
||||
)
|
||||
parser.add_argument("--num-gpu-blocks", type=int, default=100000)
|
||||
parser.add_argument(
|
||||
"--num-iteration",
|
||||
type=int,
|
||||
default=1000,
|
||||
help="Number of iterations to run to stabilize final data readings",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--allocate-blocks",
|
||||
type=int,
|
||||
nargs="*",
|
||||
default=[10, 50, 100, 500, 1000],
|
||||
help="Number of blocks to allocate",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
main(args)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
invoke_main() # pragma: no cover
|
||||
@ -19,7 +19,6 @@ import logging
|
||||
import random
|
||||
from abc import ABC, abstractmethod
|
||||
from collections.abc import Mapping
|
||||
from copy import deepcopy
|
||||
from dataclasses import dataclass
|
||||
from functools import cache
|
||||
from io import BytesIO
|
||||
@ -53,9 +52,8 @@ class SampleRequest:
|
||||
prompt: Union[str, Any]
|
||||
prompt_len: int
|
||||
expected_output_len: int
|
||||
multi_modal_data: Optional[Union[MultiModalDataDict, dict, list[dict]]] = None
|
||||
multi_modal_data: Optional[Union[MultiModalDataDict, dict]] = None
|
||||
lora_request: Optional[LoRARequest] = None
|
||||
request_id: Optional[str] = None
|
||||
|
||||
|
||||
# -----------------------------------------------------------------------------
|
||||
@ -157,10 +155,7 @@ class BenchmarkDataset(ABC):
|
||||
|
||||
@abstractmethod
|
||||
def sample(
|
||||
self,
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
num_requests: int,
|
||||
request_id_prefix: str = "",
|
||||
self, tokenizer: PreTrainedTokenizerBase, num_requests: int
|
||||
) -> list[SampleRequest]:
|
||||
"""
|
||||
Abstract method to generate sample requests from the dataset.
|
||||
@ -172,7 +167,6 @@ class BenchmarkDataset(ABC):
|
||||
tokenizer (PreTrainedTokenizerBase): The tokenizer to be used
|
||||
for processing the dataset's text.
|
||||
num_requests (int): The number of sample requests to generate.
|
||||
request_id_prefix (str) The prefix of request_id.
|
||||
|
||||
Returns:
|
||||
list[SampleRequest]: A list of sample requests generated from the
|
||||
@ -181,10 +175,7 @@ class BenchmarkDataset(ABC):
|
||||
raise NotImplementedError("sample must be implemented in subclasses.")
|
||||
|
||||
def maybe_oversample_requests(
|
||||
self,
|
||||
requests: list[SampleRequest],
|
||||
num_requests: int,
|
||||
request_id_prefix: str = "",
|
||||
self, requests: list[SampleRequest], num_requests: int
|
||||
) -> None:
|
||||
"""
|
||||
Oversamples the list of requests if its size is less than the desired
|
||||
@ -192,18 +183,11 @@ class BenchmarkDataset(ABC):
|
||||
|
||||
Args:
|
||||
requests (List[SampleRequest]): The current list of sampled
|
||||
requests.
|
||||
num_requests (int): The target number of requests.
|
||||
request_id_prefix (str) The prefix of the request ids.
|
||||
requests. num_requests (int): The target number of requests.
|
||||
"""
|
||||
if len(requests) < num_requests:
|
||||
random.seed(self.random_seed)
|
||||
additional = deepcopy(
|
||||
random.choices(requests, k=num_requests - len(requests))
|
||||
)
|
||||
for i in range(len(additional)):
|
||||
req = additional[i]
|
||||
req.request_id = request_id_prefix + str(len(requests) + i)
|
||||
additional = random.choices(requests, k=num_requests - len(requests))
|
||||
requests.extend(additional)
|
||||
logger.info("Oversampled requests to reach %d total samples.", num_requests)
|
||||
|
||||
@ -293,41 +277,6 @@ def process_image(image: Any) -> Mapping[str, Any]:
|
||||
)
|
||||
|
||||
|
||||
def process_video(video: Any) -> Mapping[str, Any]:
|
||||
"""
|
||||
Process a single video input and return a multimedia content dictionary.
|
||||
|
||||
Supports the following input types:
|
||||
|
||||
1. Dictionary with raw video bytes: - Expects a dict with a 'bytes' key
|
||||
containing raw video data.
|
||||
|
||||
2. String input: - Treats the string as a URL or local file path. -
|
||||
Prepends "file://" if the string doesn't start with "http://" or
|
||||
"file://". - Returns a dictionary with the image URL.
|
||||
|
||||
Raises:
|
||||
ValueError: If the input is not a supported type.
|
||||
"""
|
||||
if isinstance(video, dict) and "bytes" in video:
|
||||
video_bytes = video["bytes"]
|
||||
video_base64 = base64.b64encode(video_bytes).decode("utf-8")
|
||||
return {
|
||||
"type": "video_url",
|
||||
"video_url": {"url": f"data:video/mp4;base64,{video_base64}"},
|
||||
}
|
||||
|
||||
if isinstance(video, str):
|
||||
video_url = (
|
||||
video if video.startswith(("http://", "file://")) else f"file://{video}"
|
||||
)
|
||||
return {"type": "video_url", "video_url": {"url": video_url}}
|
||||
|
||||
raise ValueError(
|
||||
f"Invalid video input {video}. Must be a string of local path/remote url, or a dictionary with raw video bytes in the form of `{{'bytes': raw_video_bytes}}`." # noqa: E501
|
||||
)
|
||||
|
||||
|
||||
# -----------------------------------------------------------------------------
|
||||
# Random Dataset Implementation (Synthetic Data)
|
||||
# -----------------------------------------------------------------------------
|
||||
@ -354,7 +303,6 @@ class RandomDataset(BenchmarkDataset):
|
||||
range_ratio: float = DEFAULT_RANGE_RATIO,
|
||||
input_len: int = DEFAULT_INPUT_LEN,
|
||||
output_len: int = DEFAULT_OUTPUT_LEN,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list[SampleRequest]:
|
||||
# Enforce range_ratio < 1
|
||||
@ -376,9 +324,6 @@ class RandomDataset(BenchmarkDataset):
|
||||
input_low = int(real_input_len * (1 - range_ratio))
|
||||
input_high = int(real_input_len * (1 + range_ratio))
|
||||
output_low = int(output_len * (1 - range_ratio))
|
||||
# Ensure the lower bound for output length is at least 1 to prevent
|
||||
# sampling 0 tokens, which can cause request failures.
|
||||
output_low = max(output_low, 1)
|
||||
output_high = int(output_len * (1 + range_ratio))
|
||||
|
||||
# Add logging for debugging
|
||||
@ -403,7 +348,7 @@ class RandomDataset(BenchmarkDataset):
|
||||
# [6880, 6881] -> ['Ġcalls', 'here'] ->
|
||||
# [1650, 939, 486] -> ['Ġcall', 'sh', 'ere']
|
||||
# To avoid uncontrolled change of the prompt length,
|
||||
# the encoded sequence is truncated before being decoded again.
|
||||
# the encoded sequence is truncated before being decode again.
|
||||
total_input_len = prefix_len + int(input_lens[i])
|
||||
re_encoded_sequence = tokenizer.encode(prompt, add_special_tokens=False)[
|
||||
:total_input_len
|
||||
@ -415,10 +360,8 @@ class RandomDataset(BenchmarkDataset):
|
||||
prompt=prompt,
|
||||
prompt_len=total_input_len,
|
||||
expected_output_len=int(output_lens[i]),
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
|
||||
return requests
|
||||
|
||||
|
||||
@ -460,11 +403,9 @@ class ShareGPTDataset(BenchmarkDataset):
|
||||
max_loras: Optional[int] = None,
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
samples: list = []
|
||||
ind = 0
|
||||
for entry in self.data:
|
||||
if len(samples) >= num_requests:
|
||||
break
|
||||
@ -486,26 +427,17 @@ class ShareGPTDataset(BenchmarkDataset):
|
||||
skip_min_output_len_check=output_len is not None,
|
||||
):
|
||||
continue
|
||||
if image_path := entry.get("image"):
|
||||
mm_content = process_image(image_path)
|
||||
elif video_path := entry.get("video"):
|
||||
mm_content = process_video(video_path)
|
||||
else:
|
||||
mm_content = None
|
||||
if enable_multimodal_chat:
|
||||
prompt = self.apply_multimodal_chat_transformation(prompt, mm_content)
|
||||
prompt = self.apply_multimodal_chat_transformation(prompt, None)
|
||||
samples.append(
|
||||
SampleRequest(
|
||||
prompt=prompt,
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=new_output_len,
|
||||
lora_request=lora_request,
|
||||
multi_modal_data=mm_content,
|
||||
request_id=request_id_prefix + str(ind),
|
||||
)
|
||||
)
|
||||
ind += 1
|
||||
self.maybe_oversample_requests(samples, num_requests, request_id_prefix)
|
||||
self.maybe_oversample_requests(samples, num_requests)
|
||||
return samples
|
||||
|
||||
|
||||
@ -571,11 +503,10 @@ class CustomDataset(BenchmarkDataset):
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
skip_chat_template: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
sampled_requests = []
|
||||
for i, item in enumerate(self.data):
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
break
|
||||
prompt = item["prompt"]
|
||||
@ -594,12 +525,9 @@ class CustomDataset(BenchmarkDataset):
|
||||
prompt=prompt,
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
|
||||
return sampled_requests
|
||||
|
||||
@ -641,7 +569,6 @@ class SonnetDataset(BenchmarkDataset):
|
||||
input_len: int = DEFAULT_INPUT_LEN,
|
||||
output_len: int = DEFAULT_OUTPUT_LEN,
|
||||
return_prompt_formatted: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
# Calculate average token length for a poem line.
|
||||
@ -667,7 +594,6 @@ class SonnetDataset(BenchmarkDataset):
|
||||
prefix_lines = self.data[:num_prefix_lines]
|
||||
|
||||
samples = []
|
||||
ind = 0
|
||||
while len(samples) < num_requests:
|
||||
extra_lines = random.choices(
|
||||
self.data, k=num_input_lines - num_prefix_lines
|
||||
@ -678,17 +604,14 @@ class SonnetDataset(BenchmarkDataset):
|
||||
msg, add_generation_prompt=True, tokenize=False
|
||||
)
|
||||
prompt_len = len(tokenizer(prompt_formatted).input_ids)
|
||||
|
||||
if prompt_len <= input_len:
|
||||
samples.append(
|
||||
SampleRequest(
|
||||
prompt=prompt_formatted if return_prompt_formatted else prompt,
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
request_id=request_id_prefix + str(ind),
|
||||
)
|
||||
)
|
||||
ind += 1
|
||||
return samples
|
||||
|
||||
|
||||
@ -740,7 +663,6 @@ class BurstGPTDataset(BenchmarkDataset):
|
||||
num_requests: int,
|
||||
max_loras: Optional[int] = None,
|
||||
lora_path: Optional[str] = None,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list[SampleRequest]:
|
||||
samples = []
|
||||
@ -762,7 +684,6 @@ class BurstGPTDataset(BenchmarkDataset):
|
||||
prompt_len=input_len,
|
||||
expected_output_len=output_len,
|
||||
lora_request=lora_req,
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
return samples
|
||||
@ -780,7 +701,6 @@ class HuggingFaceDataset(BenchmarkDataset):
|
||||
self,
|
||||
dataset_path: str,
|
||||
dataset_split: str,
|
||||
no_stream: bool = False,
|
||||
dataset_subset: Optional[str] = None,
|
||||
**kwargs,
|
||||
) -> None:
|
||||
@ -788,7 +708,6 @@ class HuggingFaceDataset(BenchmarkDataset):
|
||||
|
||||
self.dataset_split = dataset_split
|
||||
self.dataset_subset = dataset_subset
|
||||
self.load_stream = not no_stream
|
||||
self.load_data()
|
||||
|
||||
def load_data(self) -> None:
|
||||
@ -797,7 +716,7 @@ class HuggingFaceDataset(BenchmarkDataset):
|
||||
self.dataset_path,
|
||||
name=self.dataset_subset,
|
||||
split=self.dataset_split,
|
||||
streaming=self.load_stream,
|
||||
streaming=True,
|
||||
)
|
||||
self.data = self.data.shuffle(seed=self.random_seed)
|
||||
|
||||
@ -822,14 +741,12 @@ class ConversationDataset(HuggingFaceDataset):
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
# Filter examples with at least 2 conversations
|
||||
filtered_data = self.data.filter(lambda x: len(x["conversations"]) >= 2)
|
||||
sampled_requests = []
|
||||
dynamic_output = output_len is None
|
||||
ind = 0
|
||||
|
||||
for item in filtered_data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
@ -857,13 +774,9 @@ class ConversationDataset(HuggingFaceDataset):
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
multi_modal_data=mm_content,
|
||||
request_id=request_id_prefix + str(ind),
|
||||
)
|
||||
)
|
||||
ind += 1
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
|
||||
@ -890,12 +803,11 @@ class VisionArenaDataset(HuggingFaceDataset):
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
output_len = output_len if output_len is not None else self.DEFAULT_OUTPUT_LEN
|
||||
sampled_requests = []
|
||||
for i, item in enumerate(self.data):
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
break
|
||||
parser_fn = self.SUPPORTED_DATASET_PATHS.get(self.dataset_path)
|
||||
@ -915,12 +827,9 @@ class VisionArenaDataset(HuggingFaceDataset):
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
multi_modal_data=mm_content,
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
|
||||
@ -950,18 +859,15 @@ class InstructCoderDataset(HuggingFaceDataset):
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
output_len = output_len if output_len is not None else self.DEFAULT_OUTPUT_LEN
|
||||
sampled_requests = []
|
||||
for i, item in enumerate(self.data):
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
break
|
||||
prompt = (
|
||||
f"{item['input']}\n\n{item['instruction']} Just output "
|
||||
"the code, do not include any explanation."
|
||||
)
|
||||
prompt = f"{item['input']}\n\n{item['instruction']} Just output \
|
||||
the code, do not include any explanation."
|
||||
|
||||
# apply template
|
||||
prompt = tokenizer.apply_chat_template(
|
||||
@ -975,12 +881,9 @@ class InstructCoderDataset(HuggingFaceDataset):
|
||||
prompt=prompt,
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
|
||||
@ -1010,13 +913,12 @@ class MTBenchDataset(HuggingFaceDataset):
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
enable_multimodal_chat: bool = False,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
output_len = output_len if output_len is not None else self.DEFAULT_OUTPUT_LEN
|
||||
sampled_requests = []
|
||||
|
||||
for i, item in enumerate(self.data):
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
break
|
||||
prompt = item["turns"][0]
|
||||
@ -1034,12 +936,9 @@ class MTBenchDataset(HuggingFaceDataset):
|
||||
prompt=prompt,
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
|
||||
@ -1064,12 +963,10 @@ class AIMODataset(HuggingFaceDataset):
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
sampled_requests = []
|
||||
dynamic_output = output_len is None
|
||||
ind = 0
|
||||
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
@ -1092,13 +989,9 @@ class AIMODataset(HuggingFaceDataset):
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
multi_modal_data=None,
|
||||
request_id=request_id_prefix + str(ind),
|
||||
)
|
||||
)
|
||||
ind += 1
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
|
||||
@ -1168,18 +1061,12 @@ class NextEditPredictionDataset(HuggingFaceDataset):
|
||||
"zed-industries/zeta": _format_zeta_prompt,
|
||||
}
|
||||
|
||||
def sample(
|
||||
self,
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
num_requests: int,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
):
|
||||
def sample(self, tokenizer: PreTrainedTokenizerBase, num_requests: int, **kwargs):
|
||||
formatting_prompt_func = self.MAPPING_PROMPT_FUNCS.get(self.dataset_path)
|
||||
if formatting_prompt_func is None:
|
||||
raise ValueError(f"Unsupported dataset path: {self.dataset_path}")
|
||||
samples = []
|
||||
for i, sample in enumerate(self.data):
|
||||
for sample in self.data:
|
||||
sample = formatting_prompt_func(sample)
|
||||
samples.append(
|
||||
SampleRequest(
|
||||
@ -1188,12 +1075,11 @@ class NextEditPredictionDataset(HuggingFaceDataset):
|
||||
expected_output_len=len(
|
||||
tokenizer(sample["expected_output"]).input_ids
|
||||
),
|
||||
request_id=request_id_prefix + str(i),
|
||||
)
|
||||
)
|
||||
if len(samples) >= num_requests:
|
||||
break
|
||||
self.maybe_oversample_requests(samples, num_requests, request_id_prefix)
|
||||
self.maybe_oversample_requests(samples, num_requests)
|
||||
return samples
|
||||
|
||||
|
||||
@ -1242,7 +1128,6 @@ class ASRDataset(HuggingFaceDataset):
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
num_requests: int,
|
||||
output_len: Optional[int] = None,
|
||||
request_id_prefix: str = "",
|
||||
**kwargs,
|
||||
) -> list:
|
||||
import librosa
|
||||
@ -1252,7 +1137,6 @@ class ASRDataset(HuggingFaceDataset):
|
||||
prompt_len = len(tokenizer(prompt).input_ids)
|
||||
sampled_requests = []
|
||||
skipped = 0
|
||||
ind = 0
|
||||
for item in self.data:
|
||||
if len(sampled_requests) >= num_requests:
|
||||
break
|
||||
@ -1271,10 +1155,8 @@ class ASRDataset(HuggingFaceDataset):
|
||||
prompt_len=prompt_len,
|
||||
expected_output_len=output_len,
|
||||
multi_modal_data=mm_content,
|
||||
request_id=request_id_prefix + str(ind),
|
||||
)
|
||||
)
|
||||
ind += 1
|
||||
if skipped:
|
||||
logger.warning(
|
||||
"%d samples discarded from dataset due to"
|
||||
@ -1282,7 +1164,5 @@ class ASRDataset(HuggingFaceDataset):
|
||||
" what Whisper supports.",
|
||||
skipped,
|
||||
)
|
||||
self.maybe_oversample_requests(
|
||||
sampled_requests, num_requests, request_id_prefix
|
||||
)
|
||||
self.maybe_oversample_requests(sampled_requests, num_requests)
|
||||
return sampled_requests
|
||||
|
||||
@ -11,7 +11,6 @@ from typing import Any, Optional
|
||||
|
||||
import numpy as np
|
||||
from tqdm import tqdm
|
||||
from typing_extensions import deprecated
|
||||
|
||||
import vllm.envs as envs
|
||||
from benchmark_utils import convert_to_pytorch_benchmark_format, write_to_json
|
||||
@ -35,10 +34,6 @@ def save_to_pytorch_benchmark_format(
|
||||
write_to_json(pt_file, pt_records)
|
||||
|
||||
|
||||
@deprecated(
|
||||
"benchmark_latency.py is deprecated and will be removed in a "
|
||||
"future version. Please use 'vllm bench latency' instead.",
|
||||
)
|
||||
def main(args: argparse.Namespace):
|
||||
print(args)
|
||||
|
||||
|
||||
@ -1,112 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import gc
|
||||
|
||||
import numpy as np
|
||||
from tabulate import tabulate
|
||||
|
||||
from benchmark_utils import TimeCollector
|
||||
from vllm.config import ModelConfig, SpeculativeConfig, VllmConfig
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
from vllm.v1.spec_decode.ngram_proposer import NgramProposer
|
||||
|
||||
|
||||
def main(args):
|
||||
rows = []
|
||||
for max_ngram in args.max_ngram:
|
||||
collector = TimeCollector(TimeCollector.US)
|
||||
|
||||
model_config = ModelConfig(
|
||||
model="facebook/opt-125m",
|
||||
task="generate",
|
||||
max_model_len=args.num_token + args.num_spec_token,
|
||||
tokenizer="facebook/opt-125m",
|
||||
tokenizer_mode="auto",
|
||||
dtype="auto",
|
||||
seed=None,
|
||||
trust_remote_code=False,
|
||||
)
|
||||
proposer = NgramProposer(
|
||||
vllm_config=VllmConfig(
|
||||
model_config=model_config,
|
||||
speculative_config=SpeculativeConfig(
|
||||
prompt_lookup_min=args.min_ngram,
|
||||
prompt_lookup_max=max_ngram,
|
||||
num_speculative_tokens=args.num_spec_token,
|
||||
method="ngram",
|
||||
),
|
||||
)
|
||||
)
|
||||
|
||||
# Warm up
|
||||
proposer.propose(np.random.randint(0, 20, (args.num_token,)))
|
||||
|
||||
gc.collect()
|
||||
for _ in range(args.num_iteration):
|
||||
tokens = np.random.randint(0, 20, (args.num_req, args.num_token))
|
||||
with collector:
|
||||
for i in range(args.num_req):
|
||||
proposer.propose(tokens[i, :])
|
||||
rows.append(
|
||||
[args.num_req, args.num_token, args.min_ngram, max_ngram]
|
||||
+ collector.dump_avg_max()
|
||||
)
|
||||
|
||||
print(
|
||||
tabulate(
|
||||
rows,
|
||||
headers=[
|
||||
"# Request",
|
||||
"# Token",
|
||||
"Min Ngram",
|
||||
"Max Ngram",
|
||||
"Avg (us)",
|
||||
"Max (us)",
|
||||
],
|
||||
tablefmt="grid",
|
||||
floatfmt=".3f",
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def invoke_main() -> None:
|
||||
parser = FlexibleArgumentParser(
|
||||
description="Benchmark the performance of N-gram speculative decode drafting"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-iteration",
|
||||
type=int,
|
||||
default=100,
|
||||
help="Number of iterations to run to stabilize final data readings",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-req", type=int, default=128, help="Number of requests in the batch"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-token", type=int, default=1500, help="Number of tokens for each request"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--min-ngram",
|
||||
type=int,
|
||||
default=3,
|
||||
help="Minimum n-gram to match",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--max-ngram",
|
||||
type=int,
|
||||
nargs="*",
|
||||
default=[5, 7, 10, 15, 20],
|
||||
help="Maximum n-gram to match",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-spec-token",
|
||||
type=int,
|
||||
default=3,
|
||||
help="Number of speculative tokens to generate",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
main(args)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
invoke_main() # pragma: no cover
|
||||
362
benchmarks/benchmark_one_concurrent.py
Normal file
362
benchmarks/benchmark_one_concurrent.py
Normal file
@ -0,0 +1,362 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
import argparse
|
||||
import asyncio
|
||||
import logging
|
||||
import random
|
||||
import time
|
||||
from dataclasses import dataclass
|
||||
from typing import Optional
|
||||
|
||||
import aiohttp # Import aiohttp
|
||||
import numpy as np
|
||||
from tqdm import tqdm
|
||||
|
||||
from backend_request_func import RequestFuncInput, RequestFuncOutput
|
||||
from benchmark_dataset import RandomDataset, SampleRequest
|
||||
|
||||
try:
|
||||
from vllm.transformers_utils.tokenizer import get_tokenizer
|
||||
except ImportError:
|
||||
from backend_request_func import get_tokenizer
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
@dataclass
|
||||
class BenchmarkMetrics:
|
||||
completed: int
|
||||
total_input: int
|
||||
total_output: int
|
||||
mean_ttft_ms: float
|
||||
median_ttft_ms: float
|
||||
std_ttft_ms: float
|
||||
percentiles_ttft_ms: list[tuple[float, float]]
|
||||
mean_itl_ms: float
|
||||
median_itl_ms: float
|
||||
std_itl_ms: float
|
||||
percentiles_itl_ms: list[tuple[float, float]]
|
||||
mean_e2el_ms: float
|
||||
median_e2el_ms: float
|
||||
std_e2el_ms: float
|
||||
percentiles_e2el_ms: list[tuple[float, float]]
|
||||
|
||||
|
||||
async def reset_cache(reset_url: str):
|
||||
"""Sends a POST request to reset the prefix cache."""
|
||||
logger.debug("Resetting prefix cache at %s", reset_url)
|
||||
try:
|
||||
async with (
|
||||
aiohttp.ClientSession() as session,
|
||||
session.post(reset_url) as response,
|
||||
):
|
||||
response.raise_for_status() # Raise an exception for bad status codes (4xx or 5xx)
|
||||
logger.debug("Prefix cache reset successful: %s", response.status)
|
||||
except aiohttp.ClientConnectorError as e:
|
||||
logger.error("Failed to connect to cache reset endpoint %s: %s}", reset_url, e)
|
||||
except aiohttp.ClientResponseError as e:
|
||||
logger.error(
|
||||
"Cache reset request failed with status %s: %s", e.status, e.message
|
||||
)
|
||||
except Exception as e:
|
||||
logger.error("An unexpected error occurred during cache reset: %s", e)
|
||||
|
||||
|
||||
async def sequential_benchmark(
|
||||
backend: str,
|
||||
api_url: str,
|
||||
model_id: str,
|
||||
tokenizer,
|
||||
input_requests: list[SampleRequest],
|
||||
request_func,
|
||||
selected_percentiles: list[float],
|
||||
cache_reset_url: Optional[str] = None,
|
||||
):
|
||||
"""
|
||||
Benchmark that processes requests sequentially, waiting for each to complete
|
||||
before starting the next one. Resets prefix cache between requests.
|
||||
"""
|
||||
outputs = []
|
||||
|
||||
pbar = tqdm(total=len(input_requests))
|
||||
|
||||
benchmark_start_time = time.perf_counter()
|
||||
|
||||
# Process requests sequentially
|
||||
for request in input_requests:
|
||||
prompt, prompt_len, output_len = (
|
||||
request.prompt,
|
||||
request.prompt_len,
|
||||
request.expected_output_len,
|
||||
)
|
||||
|
||||
logger.info("Sending request with len %s", request.prompt_len)
|
||||
logger.debug('Request str: "%s"', request.prompt[:50])
|
||||
request_start_time = time.perf_counter()
|
||||
|
||||
request_func_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
prompt=prompt,
|
||||
api_url=api_url,
|
||||
prompt_len=prompt_len,
|
||||
output_len=output_len,
|
||||
)
|
||||
|
||||
output = await request_func(request_func_input=request_func_input)
|
||||
|
||||
request_end_time = time.perf_counter()
|
||||
# Add timing information
|
||||
if output.success and not hasattr(output, "latency"):
|
||||
output.latency = request_end_time - request_start_time
|
||||
logger.info("Finished request with latency %.4f s", output.latency)
|
||||
|
||||
outputs.append(output)
|
||||
pbar.update(1)
|
||||
|
||||
pbar.close()
|
||||
|
||||
benchmark_duration = time.perf_counter() - benchmark_start_time
|
||||
|
||||
# Calculate metrics
|
||||
metrics = calculate_metrics(
|
||||
input_requests=input_requests,
|
||||
outputs=outputs,
|
||||
dur_s=benchmark_duration,
|
||||
tokenizer=tokenizer,
|
||||
selected_percentiles=selected_percentiles,
|
||||
)
|
||||
|
||||
print_results(metrics, benchmark_duration)
|
||||
|
||||
result = {
|
||||
"duration": benchmark_duration,
|
||||
"completed": metrics.completed,
|
||||
"total_input_tokens": metrics.total_input,
|
||||
"total_output_tokens": metrics.total_output,
|
||||
"input_lens": [request.prompt_len for request in input_requests],
|
||||
"output_lens": [
|
||||
output.output_tokens if output.success else 0 for output in outputs
|
||||
],
|
||||
"ttfts": [output.ttft for output in outputs if output.success],
|
||||
"itls": [output.itl for output in outputs if output.success],
|
||||
"generated_texts": [
|
||||
output.generated_text for output in outputs if output.success
|
||||
],
|
||||
"errors": [output.error for output in outputs if not output.success],
|
||||
}
|
||||
|
||||
# Add summary statistics
|
||||
for stat_name in ["ttft", "itl", "e2el"]:
|
||||
for metric_name in ["mean", "median", "std"]:
|
||||
result[f"{metric_name}_{stat_name}_ms"] = getattr(
|
||||
metrics, f"{metric_name}_{stat_name}_ms"
|
||||
)
|
||||
|
||||
for p, value in getattr(metrics, f"percentiles_{stat_name}_ms"):
|
||||
p_word = str(int(p)) if int(p) == p else str(p)
|
||||
result[f"p{p_word}_{stat_name}_ms"] = value
|
||||
|
||||
return result
|
||||
|
||||
|
||||
def calculate_metrics(
|
||||
input_requests: list[SampleRequest],
|
||||
outputs: list[RequestFuncOutput],
|
||||
dur_s: float,
|
||||
tokenizer,
|
||||
selected_percentiles: list[float],
|
||||
) -> BenchmarkMetrics:
|
||||
"""Calculate benchmark metrics from results."""
|
||||
total_input = 0
|
||||
completed = 0
|
||||
total_output = 0
|
||||
ttfts = []
|
||||
itls = []
|
||||
e2els = []
|
||||
|
||||
for i, output in enumerate(outputs):
|
||||
if output.success:
|
||||
output_len = output.output_tokens
|
||||
|
||||
if not output_len:
|
||||
# Use tokenizer to count output tokens if not provided
|
||||
output_len = len(
|
||||
tokenizer(output.generated_text, add_special_tokens=False).input_ids
|
||||
)
|
||||
|
||||
total_output += output_len
|
||||
total_input += input_requests[i].prompt_len
|
||||
|
||||
if hasattr(output, "ttft") and output.ttft is not None:
|
||||
ttfts.append(output.ttft)
|
||||
|
||||
if hasattr(output, "itl") and output.itl:
|
||||
# Ensure itl is a list of floats
|
||||
if isinstance(output.itl, list):
|
||||
itls.extend(output.itl)
|
||||
else:
|
||||
logger.warning(
|
||||
"Expected list for ITL but got %s. Appending as is.",
|
||||
type(output.itl),
|
||||
)
|
||||
itls.append(output.itl)
|
||||
|
||||
if hasattr(output, "latency") and output.latency is not None:
|
||||
e2els.append(output.latency)
|
||||
|
||||
completed += 1
|
||||
|
||||
return BenchmarkMetrics(
|
||||
completed=completed,
|
||||
total_input=total_input,
|
||||
total_output=total_output,
|
||||
mean_ttft_ms=np.mean(ttfts or [0]) * 1000,
|
||||
median_ttft_ms=np.median(ttfts or [0]) * 1000,
|
||||
std_ttft_ms=np.std(ttfts or [0]) * 1000,
|
||||
percentiles_ttft_ms=[
|
||||
(p, np.percentile(ttfts or [0], p) * 1000) for p in selected_percentiles
|
||||
],
|
||||
mean_itl_ms=np.mean(itls or [0]) * 1000,
|
||||
median_itl_ms=np.median(itls or [0]) * 1000,
|
||||
std_itl_ms=np.std(itls or [0]) * 1000,
|
||||
percentiles_itl_ms=[
|
||||
(p, np.percentile(itls or [0], p) * 1000) for p in selected_percentiles
|
||||
],
|
||||
mean_e2el_ms=np.mean(e2els or [0]) * 1000,
|
||||
median_e2el_ms=np.median(e2els or [0]) * 1000,
|
||||
std_e2el_ms=np.std(e2els or [0]) * 1000,
|
||||
percentiles_e2el_ms=[
|
||||
(p, np.percentile(e2els or [0], p) * 1000) for p in selected_percentiles
|
||||
],
|
||||
)
|
||||
|
||||
|
||||
def print_results(metrics: BenchmarkMetrics, benchmark_duration: float):
|
||||
"""Print benchmark results in a formatted way."""
|
||||
print("{s:{c}^{n}}".format(s=" Sequential Benchmark Result ", n=60, c="="))
|
||||
print("{:<40} {:<10}".format("Successful requests:", metrics.completed))
|
||||
print("{:<40} {:<10.2f}".format("Benchmark duration (s):", benchmark_duration))
|
||||
print("{:<40} {:<10}".format("Total input tokens:", metrics.total_input))
|
||||
print("{:<40} {:<10}".format("Total generated tokens:", metrics.total_output))
|
||||
|
||||
def print_metric_stats(metric_name, header):
|
||||
print("{s:{c}^{n}}".format(s=header, n=60, c="-"))
|
||||
print(
|
||||
"{:<40} {:<10.2f}".format(
|
||||
f"Mean {metric_name} (ms):",
|
||||
getattr(metrics, f"mean_{metric_name.lower()}_ms"),
|
||||
)
|
||||
)
|
||||
print(
|
||||
"{:<40} {:<10.2f}".format(
|
||||
f"Median {metric_name} (ms):",
|
||||
getattr(metrics, f"median_{metric_name.lower()}_ms"),
|
||||
)
|
||||
)
|
||||
|
||||
for p, value in getattr(metrics, f"percentiles_{metric_name.lower()}_ms"):
|
||||
p_word = str(int(p)) if int(p) == p else str(p)
|
||||
print("{:<40} {:<10.2f}".format(f"P{p_word} {metric_name} (ms):", value))
|
||||
|
||||
print_metric_stats("TTFT", "Time to First Token")
|
||||
print_metric_stats("ITL", "Inter-token Latency")
|
||||
print_metric_stats("E2EL", "End-to-end Latency")
|
||||
print("=" * 60)
|
||||
|
||||
|
||||
async def main_async(args):
|
||||
# Import needed functions based on your setup
|
||||
from backend_request_func import ASYNC_REQUEST_FUNCS
|
||||
|
||||
backend = args.backend
|
||||
model_id = args.model
|
||||
tokenizer_id = args.tokenizer if args.tokenizer is not None else args.model
|
||||
|
||||
# Set up API URL
|
||||
if args.base_url is not None:
|
||||
api_url = f"{args.base_url}{args.endpoint}"
|
||||
else:
|
||||
api_url = f"http://{args.host}:{args.port}{args.endpoint}"
|
||||
|
||||
# Set up Cache Reset URL
|
||||
cache_reset_url = f"http://{args.host}:{args.port}/reset_prefix_cache"
|
||||
logger.info("Prefix cache reset configured at: %s", cache_reset_url)
|
||||
|
||||
# Get tokenizer
|
||||
tokenizer = get_tokenizer(tokenizer_id, trust_remote_code=args.trust_remote_code)
|
||||
|
||||
# Get request function
|
||||
if backend in ASYNC_REQUEST_FUNCS:
|
||||
request_func = ASYNC_REQUEST_FUNCS[backend]
|
||||
else:
|
||||
raise ValueError(f"Unknown backend: {backend}")
|
||||
|
||||
input_requests = RandomDataset().sample(
|
||||
tokenizer=tokenizer,
|
||||
num_requests=args.num_requests,
|
||||
prefix_len=0,
|
||||
input_len=args.input_len,
|
||||
output_len=args.output_len,
|
||||
range_ratio=0.0,
|
||||
)
|
||||
|
||||
# Run benchmark
|
||||
result = await sequential_benchmark(
|
||||
backend=backend,
|
||||
api_url=api_url,
|
||||
model_id=model_id,
|
||||
tokenizer=tokenizer,
|
||||
input_requests=input_requests,
|
||||
request_func=request_func,
|
||||
selected_percentiles=[50, 90, 95, 99],
|
||||
cache_reset_url=cache_reset_url,
|
||||
)
|
||||
|
||||
return result
|
||||
|
||||
|
||||
def main(args):
|
||||
print(args)
|
||||
random.seed(args.seed)
|
||||
np.random.seed(args.seed)
|
||||
|
||||
asyncio.run(main_async(args))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser(description="Sequential benchmark for LLM serving")
|
||||
parser.add_argument(
|
||||
"--backend", type=str, default="vllm", help="Backend to use for requests"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--base-url",
|
||||
type=str,
|
||||
default=None,
|
||||
help="Server base URL (overrides --host and --port)",
|
||||
)
|
||||
parser.add_argument("--host", type=str, default="127.0.0.1")
|
||||
parser.add_argument("--port", type=int, default=8000)
|
||||
parser.add_argument(
|
||||
"--endpoint", type=str, default="/v1/completions", help="API endpoint"
|
||||
)
|
||||
parser.add_argument("--model", type=str, required=True, help="Name of the model")
|
||||
parser.add_argument(
|
||||
"--tokenizer", type=str, help="Name of the tokenizer (defaults to model name)"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-requests", type=int, default=100, help="Number of requests to process"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--input-len", type=int, default=128, help="Input len for generated prompts"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--output-len", type=int, default=None, help="Override output len for requests"
|
||||
)
|
||||
parser.add_argument("--seed", type=int, default=42)
|
||||
parser.add_argument(
|
||||
"--trust-remote-code",
|
||||
action="store_true",
|
||||
help="Trust remote code from HuggingFace",
|
||||
)
|
||||
|
||||
args = parser.parse_args()
|
||||
main(args)
|
||||
@ -5,7 +5,8 @@ r"""Benchmark online serving throughput.
|
||||
On the server side, run one of the following commands:
|
||||
vLLM OpenAI API server
|
||||
vllm serve <your_model> \
|
||||
--swap-space 16
|
||||
--swap-space 16 \
|
||||
--disable-log-requests
|
||||
|
||||
On the client side, run:
|
||||
python benchmarks/benchmark_serving.py \
|
||||
@ -29,7 +30,7 @@ import os
|
||||
import random
|
||||
import time
|
||||
import warnings
|
||||
from collections.abc import Iterable
|
||||
from collections.abc import AsyncGenerator, Iterable
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime
|
||||
from typing import Any, Literal, Optional
|
||||
@ -37,7 +38,6 @@ from typing import Any, Literal, Optional
|
||||
import numpy as np
|
||||
from tqdm.asyncio import tqdm
|
||||
from transformers import PreTrainedTokenizerBase
|
||||
from typing_extensions import deprecated
|
||||
|
||||
from backend_request_func import (
|
||||
ASYNC_REQUEST_FUNCS,
|
||||
@ -73,7 +73,6 @@ from benchmark_dataset import (
|
||||
VisionArenaDataset,
|
||||
)
|
||||
from benchmark_utils import convert_to_pytorch_benchmark_format, write_to_json
|
||||
from vllm.benchmarks.serve import get_request
|
||||
|
||||
MILLISECONDS_TO_SECONDS_CONVERSION = 1000
|
||||
|
||||
@ -108,6 +107,101 @@ class BenchmarkMetrics:
|
||||
percentiles_e2el_ms: list[tuple[float, float]]
|
||||
|
||||
|
||||
def _get_current_request_rate(
|
||||
ramp_up_strategy: Optional[Literal["linear", "exponential"]],
|
||||
ramp_up_start_rps: Optional[int],
|
||||
ramp_up_end_rps: Optional[int],
|
||||
request_index: int,
|
||||
total_requests: int,
|
||||
request_rate: float,
|
||||
) -> float:
|
||||
if (
|
||||
ramp_up_strategy
|
||||
and ramp_up_start_rps is not None
|
||||
and ramp_up_end_rps is not None
|
||||
):
|
||||
progress = request_index / max(total_requests - 1, 1)
|
||||
if ramp_up_strategy == "linear":
|
||||
increase = (ramp_up_end_rps - ramp_up_start_rps) * progress
|
||||
return ramp_up_start_rps + increase
|
||||
elif ramp_up_strategy == "exponential":
|
||||
ratio = ramp_up_end_rps / ramp_up_start_rps
|
||||
return ramp_up_start_rps * (ratio**progress)
|
||||
else:
|
||||
raise ValueError(f"Unknown ramp-up strategy: {ramp_up_strategy}")
|
||||
return request_rate
|
||||
|
||||
|
||||
async def get_request(
|
||||
input_requests: list[SampleRequest],
|
||||
request_rate: float,
|
||||
burstiness: float = 1.0,
|
||||
ramp_up_strategy: Optional[Literal["linear", "exponential"]] = None,
|
||||
ramp_up_start_rps: Optional[int] = None,
|
||||
ramp_up_end_rps: Optional[int] = None,
|
||||
) -> AsyncGenerator[tuple[SampleRequest, float], None]:
|
||||
"""
|
||||
Asynchronously generates requests at a specified rate
|
||||
with OPTIONAL burstiness and OPTIONAL ramp-up strategy.
|
||||
|
||||
Args:
|
||||
input_requests:
|
||||
A list of input requests, each represented as a SampleRequest.
|
||||
request_rate:
|
||||
The rate at which requests are generated (requests/s).
|
||||
burstiness (optional):
|
||||
The burstiness factor of the request generation.
|
||||
Only takes effect when request_rate is not inf.
|
||||
Default value is 1, which follows a Poisson process.
|
||||
Otherwise, the request intervals follow a gamma distribution.
|
||||
A lower burstiness value (0 < burstiness < 1) results
|
||||
in more bursty requests, while a higher burstiness value
|
||||
(burstiness > 1) results in a more uniform arrival of requests.
|
||||
ramp_up_strategy (optional):
|
||||
The ramp-up strategy. Can be "linear" or "exponential".
|
||||
If None, uses constant request rate (specified by request_rate).
|
||||
ramp_up_start_rps (optional):
|
||||
The starting request rate for ramp-up.
|
||||
ramp_up_end_rps (optional):
|
||||
The ending request rate for ramp-up.
|
||||
"""
|
||||
assert burstiness > 0, (
|
||||
f"A positive burstiness factor is expected, but given {burstiness}."
|
||||
)
|
||||
# Convert to list to get length for ramp-up calculations
|
||||
if isinstance(input_requests, Iterable) and not isinstance(input_requests, list):
|
||||
input_requests = list(input_requests)
|
||||
|
||||
total_requests = len(input_requests)
|
||||
request_index = 0
|
||||
|
||||
for request in input_requests:
|
||||
current_request_rate = _get_current_request_rate(
|
||||
ramp_up_strategy,
|
||||
ramp_up_start_rps,
|
||||
ramp_up_end_rps,
|
||||
request_index,
|
||||
total_requests,
|
||||
request_rate,
|
||||
)
|
||||
|
||||
yield request, current_request_rate
|
||||
|
||||
request_index += 1
|
||||
|
||||
if current_request_rate == float("inf"):
|
||||
# If the request rate is infinity, then we don't need to wait.
|
||||
continue
|
||||
|
||||
theta = 1.0 / (current_request_rate * burstiness)
|
||||
|
||||
# Sample the request interval from the gamma distribution.
|
||||
# If burstiness is 1, it follows exponential distribution.
|
||||
interval = np.random.gamma(shape=burstiness, scale=theta)
|
||||
# The next request will be sent after the interval.
|
||||
await asyncio.sleep(interval)
|
||||
|
||||
|
||||
def calculate_metrics(
|
||||
input_requests: list[SampleRequest],
|
||||
outputs: list[RequestFuncOutput],
|
||||
@ -263,14 +357,7 @@ async def benchmark(
|
||||
input_requests[0].multi_modal_data,
|
||||
)
|
||||
|
||||
assert (
|
||||
test_mm_content is None
|
||||
or isinstance(test_mm_content, dict)
|
||||
or (
|
||||
isinstance(test_mm_content, list)
|
||||
and all(isinstance(item, dict) for item in test_mm_content)
|
||||
)
|
||||
), "multi_modal_data must be a dict or list[dict]"
|
||||
assert test_mm_content is None or isinstance(test_mm_content, dict)
|
||||
test_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
model_name=model_name,
|
||||
@ -375,12 +462,11 @@ async def benchmark(
|
||||
rps_change_events.append({"rps": rps_val, "timestamp": timestamp})
|
||||
last_int_rps = current_int_rps
|
||||
|
||||
prompt, prompt_len, output_len, mm_content, request_id = (
|
||||
prompt, prompt_len, output_len, mm_content = (
|
||||
request.prompt,
|
||||
request.prompt_len,
|
||||
request.expected_output_len,
|
||||
request.multi_modal_data,
|
||||
request.request_id,
|
||||
)
|
||||
req_model_id, req_model_name = model_id, model_name
|
||||
if lora_modules:
|
||||
@ -398,12 +484,25 @@ async def benchmark(
|
||||
multi_modal_content=mm_content,
|
||||
ignore_eos=ignore_eos,
|
||||
extra_body=extra_body,
|
||||
request_id=request_id,
|
||||
)
|
||||
task = limited_request_func(request_func_input=request_func_input, pbar=pbar)
|
||||
tasks.append(asyncio.create_task(task))
|
||||
outputs: list[RequestFuncOutput] = await asyncio.gather(*tasks)
|
||||
|
||||
if profile:
|
||||
print("Stopping profiler...")
|
||||
profile_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
prompt=test_prompt,
|
||||
api_url=base_url + "/stop_profile",
|
||||
prompt_len=test_prompt_len,
|
||||
output_len=test_output_len,
|
||||
logprobs=logprobs,
|
||||
)
|
||||
profile_output = await request_func(request_func_input=profile_input)
|
||||
if profile_output.success:
|
||||
print("Profiler stopped")
|
||||
|
||||
if pbar is not None:
|
||||
pbar.close()
|
||||
|
||||
@ -421,10 +520,6 @@ async def benchmark(
|
||||
|
||||
print("{s:{c}^{n}}".format(s=" Serving Benchmark Result ", n=50, c="="))
|
||||
print("{:<40} {:<10}".format("Successful requests:", metrics.completed))
|
||||
if max_concurrency is not None:
|
||||
print("{:<40} {:<10}".format("Maximum request concurrency:", max_concurrency))
|
||||
if request_rate != float("inf"):
|
||||
print("{:<40} {:<10.2f}".format("Request rate configured (RPS):", request_rate))
|
||||
print("{:<40} {:<10.2f}".format("Benchmark duration (s):", benchmark_duration))
|
||||
print("{:<40} {:<10}".format("Total input tokens:", metrics.total_input))
|
||||
print("{:<40} {:<10}".format("Total generated tokens:", metrics.total_output))
|
||||
@ -516,20 +611,6 @@ async def benchmark(
|
||||
|
||||
print("=" * 50)
|
||||
|
||||
if profile:
|
||||
print("Stopping profiler...")
|
||||
profile_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
prompt=test_prompt,
|
||||
api_url=base_url + "/stop_profile",
|
||||
prompt_len=test_prompt_len,
|
||||
output_len=test_output_len,
|
||||
logprobs=logprobs,
|
||||
)
|
||||
profile_output = await request_func(request_func_input=profile_input)
|
||||
if profile_output.success:
|
||||
print("Profiler stopped")
|
||||
|
||||
return result
|
||||
|
||||
|
||||
@ -606,10 +687,6 @@ def save_to_pytorch_benchmark_format(
|
||||
write_to_json(pt_file, pt_records)
|
||||
|
||||
|
||||
@deprecated(
|
||||
"benchmark_serving.py is deprecated and will be removed in a future "
|
||||
"version. Please use 'vllm bench serve' instead.",
|
||||
)
|
||||
def main(args: argparse.Namespace):
|
||||
print(args)
|
||||
random.seed(args.seed)
|
||||
@ -667,7 +744,6 @@ def main(args: argparse.Namespace):
|
||||
tokenizer=tokenizer,
|
||||
output_len=args.custom_output_len,
|
||||
skip_chat_template=args.custom_skip_chat_template,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
)
|
||||
|
||||
elif args.dataset_name == "sonnet":
|
||||
@ -681,7 +757,6 @@ def main(args: argparse.Namespace):
|
||||
prefix_len=args.sonnet_prefix_len,
|
||||
tokenizer=tokenizer,
|
||||
return_prompt_formatted=False,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
)
|
||||
else:
|
||||
assert tokenizer.chat_template or tokenizer.default_chat_template, (
|
||||
@ -694,7 +769,6 @@ def main(args: argparse.Namespace):
|
||||
prefix_len=args.sonnet_prefix_len,
|
||||
tokenizer=tokenizer,
|
||||
return_prompt_formatted=True,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
)
|
||||
|
||||
elif args.dataset_name == "hf":
|
||||
@ -751,12 +825,10 @@ def main(args: argparse.Namespace):
|
||||
dataset_subset=args.hf_subset,
|
||||
dataset_split=args.hf_split,
|
||||
random_seed=args.seed,
|
||||
no_stream=args.no_stream,
|
||||
).sample(
|
||||
num_requests=args.num_prompts,
|
||||
tokenizer=tokenizer,
|
||||
output_len=args.hf_output_len,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
)
|
||||
|
||||
else:
|
||||
@ -768,15 +840,10 @@ def main(args: argparse.Namespace):
|
||||
tokenizer=tokenizer,
|
||||
num_requests=args.num_prompts,
|
||||
output_len=args.sharegpt_output_len,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
),
|
||||
"burstgpt": lambda: BurstGPTDataset(
|
||||
random_seed=args.seed, dataset_path=args.dataset_path
|
||||
).sample(
|
||||
tokenizer=tokenizer,
|
||||
num_requests=args.num_prompts,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
),
|
||||
).sample(tokenizer=tokenizer, num_requests=args.num_prompts),
|
||||
"random": lambda: RandomDataset(dataset_path=args.dataset_path).sample(
|
||||
tokenizer=tokenizer,
|
||||
num_requests=args.num_prompts,
|
||||
@ -784,7 +851,6 @@ def main(args: argparse.Namespace):
|
||||
input_len=args.random_input_len,
|
||||
output_len=args.random_output_len,
|
||||
range_ratio=args.random_range_ratio,
|
||||
request_id_prefix=args.request_id_prefix,
|
||||
),
|
||||
}
|
||||
|
||||
@ -967,11 +1033,6 @@ def create_argument_parser():
|
||||
help="Path to the sharegpt/sonnet dataset. "
|
||||
"Or the huggingface dataset ID if using HF dataset.",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--no-stream",
|
||||
action="store_true",
|
||||
help="Do not load the dataset in streaming mode.",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--max-concurrency",
|
||||
type=int,
|
||||
@ -1104,7 +1165,7 @@ def create_argument_parser():
|
||||
"--percentile-metrics",
|
||||
type=str,
|
||||
default="ttft,tpot,itl",
|
||||
help="Comma-separated list of selected metrics to report percentiles. "
|
||||
help="Comma-separated list of selected metrics to report percentils. "
|
||||
"This argument specifies the metrics to report percentiles. "
|
||||
'Allowed metric names are "ttft", "tpot", "itl", "e2el". '
|
||||
'Default value is "ttft,tpot,itl".',
|
||||
@ -1130,13 +1191,6 @@ def create_argument_parser():
|
||||
"goodput, refer to DistServe paper: https://arxiv.org/pdf/2401.09670 "
|
||||
"and the blog: https://hao-ai-lab.github.io/blogs/distserve",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--request-id-prefix",
|
||||
type=str,
|
||||
required=False,
|
||||
default="benchmark-serving",
|
||||
help="Specify the prefix of request id.",
|
||||
)
|
||||
|
||||
# group for dataset specific arguments
|
||||
custom_group = parser.add_argument_group("custom dataset options")
|
||||
|
||||
@ -4,7 +4,7 @@ r"""Benchmark online serving throughput with structured outputs.
|
||||
|
||||
On the server side, run one of the following commands:
|
||||
(vLLM OpenAI API server)
|
||||
vllm serve <your_model>
|
||||
vllm serve <your_model> --disable-log-requests
|
||||
|
||||
On the client side, run:
|
||||
python benchmarks/benchmark_serving_structured_output.py \
|
||||
@ -538,6 +538,20 @@ async def benchmark(
|
||||
)
|
||||
outputs: list[RequestFuncOutput] = await asyncio.gather(*tasks)
|
||||
|
||||
if profile:
|
||||
print("Stopping profiler...")
|
||||
profile_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
prompt=test_request.prompt,
|
||||
api_url=base_url + "/stop_profile",
|
||||
prompt_len=test_request.prompt_len,
|
||||
output_len=test_request.expected_output_len,
|
||||
extra_body={test_request.structure_type: test_request.schema},
|
||||
)
|
||||
profile_output = await request_func(request_func_input=profile_input)
|
||||
if profile_output.success:
|
||||
print("Profiler stopped")
|
||||
|
||||
if pbar is not None:
|
||||
pbar.close()
|
||||
|
||||
@ -555,10 +569,6 @@ async def benchmark(
|
||||
|
||||
print("{s:{c}^{n}}".format(s=" Serving Benchmark Result ", n=50, c="="))
|
||||
print("{:<40} {:<10}".format("Successful requests:", metrics.completed))
|
||||
if max_concurrency is not None:
|
||||
print("{:<40} {:<10}".format("Maximum request concurrency:", max_concurrency))
|
||||
if request_rate != float("inf"):
|
||||
print("{:<40} {:<10.2f}".format("Request rate configured (RPS):", request_rate))
|
||||
print("{:<40} {:<10.2f}".format("Benchmark duration (s):", benchmark_duration))
|
||||
print("{:<40} {:<10}".format("Total input tokens:", metrics.total_input))
|
||||
print("{:<40} {:<10}".format("Total generated tokens:", metrics.total_output))
|
||||
@ -656,20 +666,6 @@ async def benchmark(
|
||||
|
||||
print("=" * 50)
|
||||
|
||||
if profile:
|
||||
print("Stopping profiler...")
|
||||
profile_input = RequestFuncInput(
|
||||
model=model_id,
|
||||
prompt=test_request.prompt,
|
||||
api_url=base_url + "/stop_profile",
|
||||
prompt_len=test_request.prompt_len,
|
||||
output_len=test_request.expected_output_len,
|
||||
extra_body={test_request.structure_type: test_request.schema},
|
||||
)
|
||||
profile_output = await request_func(request_func_input=profile_input)
|
||||
if profile_output.success:
|
||||
print("Profiler stopped")
|
||||
|
||||
return result, ret
|
||||
|
||||
|
||||
@ -998,7 +994,7 @@ def create_argument_parser():
|
||||
"--percentile-metrics",
|
||||
type=str,
|
||||
default="ttft,tpot,itl",
|
||||
help="Comma-separated list of selected metrics to report percentiles. "
|
||||
help="Comma-separated list of selected metrics to report percentils. "
|
||||
"This argument specifies the metrics to report percentiles. "
|
||||
'Allowed metric names are "ttft", "tpot", "itl", "e2el". '
|
||||
'Default value is "ttft,tpot,itl".',
|
||||
|
||||
@ -15,7 +15,6 @@ import torch
|
||||
import uvloop
|
||||
from tqdm import tqdm
|
||||
from transformers import AutoModelForCausalLM, AutoTokenizer, PreTrainedTokenizerBase
|
||||
from typing_extensions import deprecated
|
||||
|
||||
from benchmark_dataset import (
|
||||
AIMODataset,
|
||||
@ -96,6 +95,7 @@ def run_vllm(
|
||||
end = time.perf_counter()
|
||||
else:
|
||||
assert lora_requests is None, "BeamSearch API does not support LoRA"
|
||||
prompts = [request.prompt for request in requests]
|
||||
# output_len should be the same for all requests.
|
||||
output_len = requests[0].expected_output_len
|
||||
for request in requests:
|
||||
@ -167,8 +167,7 @@ async def run_vllm_async(
|
||||
from vllm import SamplingParams
|
||||
|
||||
async with build_async_engine_client_from_engine_args(
|
||||
engine_args,
|
||||
disable_frontend_multiprocessing=disable_frontend_multiprocessing,
|
||||
engine_args, disable_frontend_multiprocessing
|
||||
) as llm:
|
||||
model_config = await llm.get_model_config()
|
||||
assert all(
|
||||
@ -357,7 +356,6 @@ def get_requests(args, tokenizer):
|
||||
elif args.dataset_name == "burstgpt":
|
||||
dataset_cls = BurstGPTDataset
|
||||
elif args.dataset_name == "hf":
|
||||
common_kwargs["no_stream"] = args.no_stream
|
||||
if args.dataset_path in VisionArenaDataset.SUPPORTED_DATASET_PATHS:
|
||||
dataset_cls = VisionArenaDataset
|
||||
common_kwargs["dataset_subset"] = None
|
||||
@ -382,10 +380,6 @@ def get_requests(args, tokenizer):
|
||||
return dataset_cls(**common_kwargs).sample(**sample_kwargs)
|
||||
|
||||
|
||||
@deprecated(
|
||||
"benchmark_throughput.py is deprecated and will be removed in a "
|
||||
"future version. Please use 'vllm bench throughput' instead.",
|
||||
)
|
||||
def main(args: argparse.Namespace):
|
||||
if args.seed is None:
|
||||
args.seed = 0
|
||||
@ -596,8 +590,8 @@ def validate_args(args):
|
||||
# https://github.com/vllm-project/vllm/issues/16222
|
||||
if args.data_parallel_size > 1:
|
||||
raise ValueError(
|
||||
"Data parallel is not supported in offline benchmark, "
|
||||
"please use benchmark serving instead"
|
||||
"Data parallel is not supported in offline benchmark, \
|
||||
please use benchmark serving instead"
|
||||
)
|
||||
|
||||
|
||||
@ -616,11 +610,6 @@ def create_argument_parser():
|
||||
help="Name of the dataset to benchmark on.",
|
||||
default="sharegpt",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--no-stream",
|
||||
action="store_true",
|
||||
help="Do not load the dataset in streaming mode.",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--dataset",
|
||||
type=str,
|
||||
@ -719,7 +708,7 @@ def create_argument_parser():
|
||||
"[length * (1 - range_ratio), length * (1 + range_ratio)].",
|
||||
)
|
||||
|
||||
# hf dataset
|
||||
# hf dtaset
|
||||
parser.add_argument(
|
||||
"--hf-subset", type=str, default=None, help="Subset of the HF dataset."
|
||||
)
|
||||
|
||||
@ -1,12 +1,11 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import math
|
||||
import os
|
||||
import time
|
||||
from types import TracebackType
|
||||
from typing import Any, Optional, Union
|
||||
from typing import Any
|
||||
|
||||
|
||||
def convert_to_pytorch_benchmark_format(
|
||||
@ -73,53 +72,3 @@ def write_to_json(filename: str, records: list) -> None:
|
||||
cls=InfEncoder,
|
||||
default=lambda o: f"<{type(o).__name__} object is not JSON serializable>",
|
||||
)
|
||||
|
||||
|
||||
# Collect time and generate time metrics
|
||||
#
|
||||
# Example Usage:
|
||||
# collector = TimeCollector(TimeCollector.US)
|
||||
# for _ in range(total_iteration):
|
||||
# with collector:
|
||||
# ...
|
||||
# collector.dump_avg_max()
|
||||
class TimeCollector:
|
||||
NS: int = 1
|
||||
US: int = NS * 1000
|
||||
MS: int = US * 1000
|
||||
S: int = MS * 1000
|
||||
|
||||
def __init__(self, scale: int) -> None:
|
||||
self.cnt: int = 0
|
||||
self._sum: int = 0
|
||||
self._max: Optional[int] = None
|
||||
self.scale = scale
|
||||
self.start_time: int = time.monotonic_ns()
|
||||
|
||||
def collect(self, v: int) -> None:
|
||||
self.cnt += 1
|
||||
self._sum += v
|
||||
if self._max is None:
|
||||
self._max = v
|
||||
else:
|
||||
self._max = max(self._max, v)
|
||||
|
||||
def avg(self) -> Union[float, str]:
|
||||
return self._sum * 1.0 / self.cnt / self.scale if self.cnt > 0 else "N/A"
|
||||
|
||||
def max(self) -> Union[float, str]:
|
||||
return self._max / self.scale if self._max else "N/A"
|
||||
|
||||
def dump_avg_max(self) -> list[Union[float, str]]:
|
||||
return [self.avg(), self.max()]
|
||||
|
||||
def __enter__(self) -> None:
|
||||
self.start_time = time.monotonic_ns()
|
||||
|
||||
def __exit__(
|
||||
self,
|
||||
exc_type: Optional[type[BaseException]],
|
||||
exc_value: Optional[BaseException],
|
||||
exc_traceback: Optional[TracebackType],
|
||||
) -> None:
|
||||
self.collect(time.monotonic_ns() - self.start_time)
|
||||
|
||||
@ -3,7 +3,7 @@
|
||||
# benchmark the overhead of disaggregated prefill.
|
||||
# methodology:
|
||||
# - send all request to prefill vLLM instance. It will buffer KV cache.
|
||||
# - then send all request to decode instance.
|
||||
# - then send all request to decode instance.
|
||||
# - The TTFT of decode instance is the overhead.
|
||||
|
||||
set -ex
|
||||
@ -12,8 +12,6 @@ kill_gpu_processes() {
|
||||
# kill all processes on GPU.
|
||||
pgrep pt_main_thread | xargs -r kill -9
|
||||
pgrep python3 | xargs -r kill -9
|
||||
# vLLM now names the process with VLLM prefix after https://github.com/vllm-project/vllm/pull/21445
|
||||
pgrep VLLM | xargs -r kill -9
|
||||
sleep 10
|
||||
|
||||
# remove vllm config file
|
||||
@ -62,8 +60,8 @@ benchmark() {
|
||||
--max-model-len 10000 \
|
||||
--gpu-memory-utilization 0.6 \
|
||||
--kv-transfer-config \
|
||||
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
|
||||
'{"kv_connector":"PyNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
|
||||
|
||||
CUDA_VISIBLE_DEVICES=1 python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
@ -72,44 +70,44 @@ benchmark() {
|
||||
--max-model-len 10000 \
|
||||
--gpu-memory-utilization 0.6 \
|
||||
--kv-transfer-config \
|
||||
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
'{"kv_connector":"PyNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
|
||||
wait_for_server 8100
|
||||
wait_for_server 8200
|
||||
|
||||
# let the prefill instance finish prefill
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8100 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename disagg_prefill_tp1.json \
|
||||
--request-rate "inf"
|
||||
python3 ../benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8100 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename disagg_prefill_tp1.json \
|
||||
--request-rate "inf"
|
||||
|
||||
|
||||
# send the request to decode.
|
||||
# The TTFT of this command will be the overhead of disagg prefill impl.
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8200 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename disagg_prefill_tp1_overhead.json \
|
||||
--request-rate "$qps"
|
||||
python3 ../benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8200 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename disagg_prefill_tp1_overhead.json \
|
||||
--request-rate "$qps"
|
||||
kill_gpu_processes
|
||||
|
||||
}
|
||||
|
||||
@ -18,8 +18,6 @@ kill_gpu_processes() {
|
||||
# kill all processes on GPU.
|
||||
pgrep pt_main_thread | xargs -r kill -9
|
||||
pgrep python3 | xargs -r kill -9
|
||||
# vLLM now names the process with VLLM prefix after https://github.com/vllm-project/vllm/pull/21445
|
||||
pgrep VLLM | xargs -r kill -9
|
||||
for port in 8000 8100 8200; do lsof -t -i:$port | xargs -r kill -9; done
|
||||
sleep 1
|
||||
}
|
||||
@ -60,7 +58,7 @@ launch_chunked_prefill() {
|
||||
|
||||
|
||||
launch_disagg_prefill() {
|
||||
model="meta-llama/Meta-Llama-3.1-8B-Instruct"
|
||||
model="meta-llama/Meta-Llama-3.1-8B-Instruct"
|
||||
# disagg prefill
|
||||
CUDA_VISIBLE_DEVICES=0 python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
@ -69,7 +67,7 @@ launch_disagg_prefill() {
|
||||
--max-model-len 10000 \
|
||||
--gpu-memory-utilization 0.6 \
|
||||
--kv-transfer-config \
|
||||
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
'{"kv_connector":"PyNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
|
||||
CUDA_VISIBLE_DEVICES=1 python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
@ -78,7 +76,7 @@ launch_disagg_prefill() {
|
||||
--max-model-len 10000 \
|
||||
--gpu-memory-utilization 0.6 \
|
||||
--kv-transfer-config \
|
||||
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
'{"kv_connector":"PyNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
|
||||
|
||||
wait_for_server 8100
|
||||
wait_for_server 8200
|
||||
@ -99,20 +97,20 @@ benchmark() {
|
||||
output_len=$2
|
||||
tag=$3
|
||||
|
||||
vllm bench serve \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8000 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename "$tag"-qps-"$qps".json \
|
||||
--request-rate "$qps"
|
||||
python3 ../benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model $model \
|
||||
--dataset-name $dataset_name \
|
||||
--dataset-path $dataset_path \
|
||||
--sonnet-input-len $input_len \
|
||||
--sonnet-output-len "$output_len" \
|
||||
--sonnet-prefix-len $prefix_len \
|
||||
--num-prompts $num_prompts \
|
||||
--port 8000 \
|
||||
--save-result \
|
||||
--result-dir $results_folder \
|
||||
--result-filename "$tag"-qps-"$qps".json \
|
||||
--request-rate "$qps"
|
||||
|
||||
sleep 2
|
||||
}
|
||||
|
||||
@ -1,199 +1,63 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import argparse
|
||||
import asyncio
|
||||
import logging
|
||||
import os
|
||||
|
||||
import aiohttp
|
||||
from quart import Quart, Response, make_response, request
|
||||
from rate_limiter import RateLimiter
|
||||
from request_queue import RequestQueue
|
||||
from quart import Quart, make_response, request
|
||||
|
||||
# Configure logging
|
||||
logging.basicConfig(level=logging.INFO)
|
||||
logger = logging.getLogger(__name__)
|
||||
AIOHTTP_TIMEOUT = aiohttp.ClientTimeout(total=6 * 60 * 60)
|
||||
|
||||
app = Quart(__name__)
|
||||
|
||||
|
||||
def parse_args():
|
||||
"""parse command line arguments"""
|
||||
parser = argparse.ArgumentParser(description="vLLM P/D disaggregation proxy server")
|
||||
|
||||
# Add args
|
||||
parser.add_argument(
|
||||
"--timeout",
|
||||
type=float,
|
||||
default=300,
|
||||
help="Timeout for backend service requests in seconds (default: 300)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--max-concurrent",
|
||||
type=int,
|
||||
default=100,
|
||||
help="Maximum concurrent requests to backend services (default: 100)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--queue-size",
|
||||
type=int,
|
||||
default=500,
|
||||
help="Maximum number of requests in the queue (default: 500)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--rate-limit",
|
||||
type=int,
|
||||
default=40,
|
||||
help="Maximum requests per second (default: 40)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--port",
|
||||
type=int,
|
||||
default=8000,
|
||||
help="Port to run the server on (default: 8000)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--prefill-url",
|
||||
type=str,
|
||||
default="http://localhost:8100/v1/completions",
|
||||
help="Prefill service endpoint URL",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--decode-url",
|
||||
type=str,
|
||||
default="http://localhost:8200/v1/completions",
|
||||
help="Decode service endpoint URL",
|
||||
)
|
||||
|
||||
return parser.parse_args()
|
||||
|
||||
|
||||
def main():
|
||||
"""parse command line arguments"""
|
||||
args = parse_args()
|
||||
|
||||
# Initialize configuration using command line parameters
|
||||
AIOHTTP_TIMEOUT = aiohttp.ClientTimeout(total=args.timeout)
|
||||
MAX_CONCURRENT_REQUESTS = args.max_concurrent
|
||||
REQUEST_QUEUE_SIZE = args.queue_size
|
||||
RATE_LIMIT = args.rate_limit
|
||||
PREFILL_SERVICE_URL = args.prefill_url
|
||||
DECODE_SERVICE_URL = args.decode_url
|
||||
PORT = args.port
|
||||
|
||||
app = Quart(__name__)
|
||||
|
||||
# Initialize the rate limiter and request queue
|
||||
rate_limiter = RateLimiter(RATE_LIMIT)
|
||||
request_queue = RequestQueue(MAX_CONCURRENT_REQUESTS, REQUEST_QUEUE_SIZE)
|
||||
|
||||
# Attach the configuration object to the application instance
|
||||
app.config.update(
|
||||
{
|
||||
"AIOHTTP_TIMEOUT": AIOHTTP_TIMEOUT,
|
||||
"rate_limiter": rate_limiter,
|
||||
"request_queue": request_queue,
|
||||
"PREFILL_SERVICE_URL": PREFILL_SERVICE_URL,
|
||||
"DECODE_SERVICE_URL": DECODE_SERVICE_URL,
|
||||
}
|
||||
)
|
||||
|
||||
# Start queue processing on app startup
|
||||
@app.before_serving
|
||||
async def startup():
|
||||
"""Start request processing task when app starts serving"""
|
||||
asyncio.create_task(request_queue.process())
|
||||
|
||||
async def forward_request(url, data):
|
||||
"""Forward request to backend service with rate limiting and error handling"""
|
||||
async def forward_request(url, data):
|
||||
async with aiohttp.ClientSession(timeout=AIOHTTP_TIMEOUT) as session:
|
||||
headers = {"Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}"}
|
||||
async with session.post(url=url, json=data, headers=headers) as response:
|
||||
if response.status == 200:
|
||||
# if response.headers.get('Transfer-Encoding') == 'chunked':
|
||||
if True:
|
||||
async for chunk_bytes in response.content.iter_chunked(1024):
|
||||
yield chunk_bytes
|
||||
else:
|
||||
content = await response.read()
|
||||
yield content
|
||||
|
||||
# Use rate limiter as context manager
|
||||
async with (
|
||||
rate_limiter,
|
||||
aiohttp.ClientSession(timeout=AIOHTTP_TIMEOUT) as session,
|
||||
|
||||
@app.route("/v1/completions", methods=["POST"])
|
||||
async def handle_request():
|
||||
try:
|
||||
original_request_data = await request.get_json()
|
||||
|
||||
prefill_request = original_request_data.copy()
|
||||
# change max_tokens = 1 to let it only do prefill
|
||||
prefill_request["max_tokens"] = 1
|
||||
|
||||
# finish prefill
|
||||
async for _ in forward_request(
|
||||
"http://localhost:8100/v1/completions", prefill_request
|
||||
):
|
||||
try:
|
||||
async with session.post(
|
||||
url=url, json=data, headers=headers
|
||||
) as response:
|
||||
if response.status == 200:
|
||||
# Stream response chunks
|
||||
async for chunk_bytes in response.content.iter_chunked(1024):
|
||||
yield chunk_bytes
|
||||
else:
|
||||
# Handle backend service errors
|
||||
error_text = await response.text()
|
||||
logger.error(
|
||||
"Backend service error: %s - %s",
|
||||
response.status,
|
||||
error_text,
|
||||
)
|
||||
yield b'{"error": "Backend service error"}'
|
||||
except aiohttp.ClientError as e:
|
||||
# Handle connection errors
|
||||
logger.error("Connection error to %s: %s", url, str(e))
|
||||
yield b'{"error": "Service unavailable"}'
|
||||
except asyncio.TimeoutError:
|
||||
# Handle timeout errors
|
||||
logger.error("Timeout connecting to %s", url)
|
||||
yield b'{"error": "Service timeout"}'
|
||||
continue
|
||||
|
||||
async def process_request():
|
||||
"""Process a single request through prefill and decode stages"""
|
||||
try:
|
||||
original_request_data = await request.get_json()
|
||||
# return decode
|
||||
generator = forward_request(
|
||||
"http://localhost:8200/v1/completions", original_request_data
|
||||
)
|
||||
response = await make_response(generator)
|
||||
response.timeout = None
|
||||
|
||||
# Create prefill request (max_tokens=1)
|
||||
prefill_request = original_request_data.copy()
|
||||
prefill_request["max_tokens"] = 1
|
||||
return response
|
||||
|
||||
# Execute prefill stage
|
||||
async for _ in forward_request(PREFILL_SERVICE_URL, prefill_request):
|
||||
continue
|
||||
except Exception as e:
|
||||
import sys
|
||||
import traceback
|
||||
|
||||
# Execute decode stage and stream response
|
||||
generator = forward_request(DECODE_SERVICE_URL, original_request_data)
|
||||
response = await make_response(generator)
|
||||
response.timeout = None # Disable timeout for streaming response
|
||||
return response
|
||||
|
||||
except Exception:
|
||||
logger.exception("Error processing request")
|
||||
return Response(
|
||||
response=b'{"error": "Internal server error"}',
|
||||
status=500,
|
||||
content_type="application/json",
|
||||
)
|
||||
|
||||
@app.route("/v1/completions", methods=["POST"])
|
||||
async def handle_request():
|
||||
"""Handle incoming API requests with concurrency and rate limiting"""
|
||||
# Create task for request processing
|
||||
task = asyncio.create_task(process_request())
|
||||
|
||||
# Enqueue request or reject if queue is full
|
||||
if not await request_queue.enqueue(task):
|
||||
return Response(
|
||||
response=b'{"error": "Server busy, try again later"}',
|
||||
status=503,
|
||||
content_type="application/json",
|
||||
)
|
||||
|
||||
try:
|
||||
# Return the response from the processing task
|
||||
return await task
|
||||
except asyncio.CancelledError:
|
||||
# Handle task cancellation (timeout or queue full)
|
||||
logger.warning("Request cancelled due to timeout or queue full")
|
||||
return Response(
|
||||
response=b'{"error": "Request cancelled"}',
|
||||
status=503,
|
||||
content_type="application/json",
|
||||
)
|
||||
|
||||
# Start the Quart server with host can be set to 0.0.0.0
|
||||
app.run(port=PORT)
|
||||
exc_info = sys.exc_info()
|
||||
print("Error occurred in disagg prefill proxy server")
|
||||
print(e)
|
||||
print("".join(traceback.format_exception(*exc_info)))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
app.run(port=8000)
|
||||
|
||||
@ -1,45 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import asyncio
|
||||
import time
|
||||
|
||||
|
||||
class RateLimiter:
|
||||
"""Token bucket rate limiter implementation"""
|
||||
|
||||
def __init__(self, rate_limit):
|
||||
self.rate_limit = rate_limit # Requests per second
|
||||
self.num_available_tokens = rate_limit # Available tokens
|
||||
self.last_refill = time.monotonic() # Last token refill time
|
||||
self.lock = asyncio.Lock() # Synchronization lock
|
||||
|
||||
async def acquire(self):
|
||||
"""Acquire a token from the rate limiter"""
|
||||
while True:
|
||||
async with self.lock:
|
||||
current_time = time.monotonic()
|
||||
elapsed = current_time - self.last_refill
|
||||
|
||||
# Refill num_available_tokens if more than 1 second has passed
|
||||
if elapsed > 1.0:
|
||||
self.num_available_tokens = self.rate_limit
|
||||
self.last_refill = current_time
|
||||
|
||||
# Check if num_available_tokens are available
|
||||
if self.num_available_tokens > 0:
|
||||
self.num_available_tokens -= 1
|
||||
return True
|
||||
|
||||
# Calculate wait time if no num_available_tokens available
|
||||
wait_time = 1.0 - elapsed
|
||||
await asyncio.sleep(wait_time)
|
||||
|
||||
async def __aenter__(self):
|
||||
"""Enter async context manager - acquire token"""
|
||||
await self.acquire()
|
||||
return self
|
||||
|
||||
async def __aexit__(self, exc_type, exc_value, traceback):
|
||||
"""Exit async context manager - no cleanup needed"""
|
||||
pass
|
||||
@ -1,39 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import asyncio
|
||||
from collections import deque
|
||||
|
||||
|
||||
class RequestQueue:
|
||||
"""Request queue manager with concurrency control"""
|
||||
|
||||
def __init__(self, max_concurrent, max_queue_size):
|
||||
# Maximum concurrent requests
|
||||
self.max_concurrent = max_concurrent
|
||||
self.max_queue_size = max_queue_size # Maximum queue size
|
||||
# Concurrency control
|
||||
self.semaphore = asyncio.Semaphore(max_concurrent)
|
||||
self.queue = deque() # Request queue
|
||||
self.queue_size = 0 # Current queue size
|
||||
self.lock = asyncio.Lock() # Sync queue Lock
|
||||
|
||||
async def enqueue(self, task):
|
||||
"""Add a request task to the queue"""
|
||||
async with self.lock:
|
||||
if self.queue_size >= self.max_queue_size:
|
||||
return False
|
||||
|
||||
self.queue.append(task)
|
||||
self.queue_size += 1
|
||||
return True
|
||||
|
||||
async def process(self):
|
||||
"""Process queued requests using semaphore for concurrency control"""
|
||||
while True:
|
||||
if self.queue:
|
||||
async with self.semaphore, self.lock:
|
||||
task = self.queue.popleft()
|
||||
self.queue_size -= 1
|
||||
await task
|
||||
await asyncio.sleep(0.01) # Yield control to event loop
|
||||
@ -1,114 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import torch
|
||||
|
||||
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
|
||||
w8a8_block_fp8_matmul,
|
||||
)
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.triton_utils import triton as vllm_triton
|
||||
|
||||
assert current_platform.is_cuda(), (
|
||||
"Only support benchmarking w8a8 block fp8 kernel on CUDA device."
|
||||
)
|
||||
|
||||
# DeepSeek-V3 weight shapes
|
||||
DEEPSEEK_V3_SHAPES = [
|
||||
(512 + 64, 7168),
|
||||
(2112, 7168),
|
||||
((128 + 64) * 128, 7168),
|
||||
(128 * (128 + 128), 512),
|
||||
(7168, 16384),
|
||||
(7168, 18432),
|
||||
(18432 * 2, 7168),
|
||||
(24576, 1536),
|
||||
(12288, 7168),
|
||||
(4096, 7168),
|
||||
(7168, 2048),
|
||||
]
|
||||
|
||||
|
||||
def build_w8a8_block_fp8_runner(M, N, K, block_size, device):
|
||||
"""Build runner function for w8a8 block fp8 matmul."""
|
||||
factor_for_scale = 1e-2
|
||||
|
||||
fp8_info = torch.finfo(torch.float8_e4m3fn)
|
||||
fp8_max, fp8_min = fp8_info.max, fp8_info.min
|
||||
|
||||
# Create random FP8 tensors
|
||||
A_fp32 = (torch.rand(M, K, dtype=torch.float32, device=device) - 0.5) * 2 * fp8_max
|
||||
A = A_fp32.clamp(min=fp8_min, max=fp8_max).to(torch.float8_e4m3fn)
|
||||
|
||||
B_fp32 = (torch.rand(N, K, dtype=torch.float32, device=device) - 0.5) * 2 * fp8_max
|
||||
B = B_fp32.clamp(min=fp8_min, max=fp8_max).to(torch.float8_e4m3fn)
|
||||
|
||||
# Create scales
|
||||
block_n, block_k = block_size[0], block_size[1]
|
||||
n_tiles = (N + block_n - 1) // block_n
|
||||
k_tiles = (K + block_k - 1) // block_k
|
||||
|
||||
As = torch.rand(M, k_tiles, dtype=torch.float32, device=device) * factor_for_scale
|
||||
Bs = (
|
||||
torch.rand(n_tiles, k_tiles, dtype=torch.float32, device=device)
|
||||
* factor_for_scale
|
||||
)
|
||||
|
||||
def run():
|
||||
return w8a8_block_fp8_matmul(A, B, As, Bs, block_size, torch.bfloat16)
|
||||
|
||||
return run
|
||||
|
||||
|
||||
@vllm_triton.testing.perf_report(
|
||||
vllm_triton.testing.Benchmark(
|
||||
x_names=["batch_size"],
|
||||
x_vals=[1, 16, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384],
|
||||
x_log=False,
|
||||
line_arg="provider",
|
||||
line_vals=["torch-bf16", "w8a8-block-fp8"],
|
||||
line_names=["torch-bf16", "w8a8-block-fp8"],
|
||||
ylabel="TFLOP/s (larger is better)",
|
||||
plot_name="BF16 vs W8A8 Block FP8 GEMMs",
|
||||
args={},
|
||||
)
|
||||
)
|
||||
def benchmark_tflops(batch_size, provider, N, K, block_size=(128, 128)):
|
||||
M = batch_size
|
||||
device = "cuda"
|
||||
|
||||
quantiles = [0.5, 0.2, 0.8]
|
||||
|
||||
if provider == "torch-bf16":
|
||||
a = torch.randn((M, K), device=device, dtype=torch.bfloat16)
|
||||
b = torch.randn((N, K), device=device, dtype=torch.bfloat16)
|
||||
ms, min_ms, max_ms = vllm_triton.testing.do_bench_cudagraph(
|
||||
lambda: torch.nn.functional.linear(a, b), quantiles=quantiles
|
||||
)
|
||||
else: # w8a8-block-fp8
|
||||
run_w8a8 = build_w8a8_block_fp8_runner(M, N, K, block_size, device)
|
||||
ms, min_ms, max_ms = vllm_triton.testing.do_bench_cudagraph(
|
||||
lambda: run_w8a8(), quantiles=quantiles
|
||||
)
|
||||
|
||||
to_tflops = lambda t_ms: (2 * M * N * K) * 1e-12 / (t_ms * 1e-3)
|
||||
return to_tflops(ms), to_tflops(max_ms), to_tflops(min_ms)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
block_size = (128, 128)
|
||||
|
||||
for N, K in DEEPSEEK_V3_SHAPES:
|
||||
print(f"\nBenchmarking DeepSeek-V3, N={N} K={K}")
|
||||
|
||||
print(f"TFLOP/s comparison (block_size={block_size}):")
|
||||
benchmark_tflops.run(
|
||||
print_data=True,
|
||||
# show_plots=False,
|
||||
# save_path=f"bench_w8a8_block_fp8_tflops_n{N}_k{K}",
|
||||
N=N,
|
||||
K=K,
|
||||
block_size=block_size,
|
||||
)
|
||||
|
||||
print("\nBenchmark finished!")
|
||||
@ -1,5 +1,4 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import argparse
|
||||
import copy
|
||||
import itertools
|
||||
|
||||
@ -1,141 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import argparse
|
||||
import copy
|
||||
import itertools
|
||||
|
||||
import torch
|
||||
from weight_shapes import WEIGHT_SHAPES
|
||||
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.scalar_type import scalar_types
|
||||
from vllm.triton_utils import triton
|
||||
|
||||
if not current_platform.has_device_capability(100):
|
||||
raise RuntimeError("NVFP4 requires compute capability of 10.0 (Blackwell)")
|
||||
|
||||
|
||||
FLOAT4_E2M1_MAX = scalar_types.float4_e2m1f.max()
|
||||
FLOAT8_E4M3_MAX = torch.finfo(torch.float8_e4m3fn).max
|
||||
|
||||
PROVIDER_CFGS = {
|
||||
"torch-bf16": dict(enabled=True),
|
||||
"nvfp4": dict(no_a_quant=False, enabled=True),
|
||||
"nvfp4-noquant": dict(no_a_quant=True, enabled=True),
|
||||
}
|
||||
|
||||
_enabled = [k for k, v in PROVIDER_CFGS.items() if v["enabled"]]
|
||||
|
||||
|
||||
def _quant_weight_nvfp4(b: torch.Tensor, device: str):
|
||||
# Compute global scale for weight
|
||||
b_amax = torch.abs(b).max().to(torch.float32)
|
||||
b_global_scale = FLOAT8_E4M3_MAX * FLOAT4_E2M1_MAX / b_amax
|
||||
b_fp4, scale_b_fp4 = ops.scaled_fp4_quant(b, b_global_scale)
|
||||
return b_fp4, scale_b_fp4, b_global_scale
|
||||
|
||||
|
||||
def build_nvfp4_runner(cfg, a, b, dtype, device):
|
||||
b_fp4, scale_b_fp4, b_global_scale = _quant_weight_nvfp4(b, device)
|
||||
|
||||
# Compute global scale for activation
|
||||
# NOTE: This is generally provided ahead-of-time by the model checkpoint.
|
||||
a_amax = torch.abs(a).max().to(torch.float32)
|
||||
a_global_scale = FLOAT8_E4M3_MAX * FLOAT4_E2M1_MAX / a_amax
|
||||
|
||||
# Alpha for the GEMM operation
|
||||
alpha = 1.0 / (a_global_scale * b_global_scale)
|
||||
|
||||
if cfg["no_a_quant"]:
|
||||
# Pre-quantize activation
|
||||
a_fp4, scale_a_fp4 = ops.scaled_fp4_quant(a, a_global_scale)
|
||||
|
||||
def run():
|
||||
return ops.cutlass_scaled_fp4_mm(
|
||||
a_fp4, b_fp4, scale_a_fp4, scale_b_fp4, alpha, dtype
|
||||
)
|
||||
|
||||
return run
|
||||
|
||||
# Quantize activation on-the-fly
|
||||
def run():
|
||||
a_fp4, scale_a_fp4 = ops.scaled_fp4_quant(a, a_global_scale)
|
||||
return ops.cutlass_scaled_fp4_mm(
|
||||
a_fp4, b_fp4, scale_a_fp4, scale_b_fp4, alpha, dtype
|
||||
)
|
||||
|
||||
return run
|
||||
|
||||
|
||||
@triton.testing.perf_report(
|
||||
triton.testing.Benchmark(
|
||||
x_names=["batch_size"],
|
||||
x_vals=[1, 16, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384],
|
||||
x_log=False,
|
||||
line_arg="provider",
|
||||
line_vals=_enabled,
|
||||
line_names=_enabled,
|
||||
ylabel="TFLOP/s (larger is better)",
|
||||
plot_name="BF16 vs NVFP4 GEMMs",
|
||||
args={},
|
||||
)
|
||||
)
|
||||
def benchmark(batch_size, provider, N, K):
|
||||
M = batch_size
|
||||
device = "cuda"
|
||||
dtype = torch.bfloat16
|
||||
|
||||
a = torch.randn((M, K), device=device, dtype=dtype)
|
||||
b = torch.randn((N, K), device=device, dtype=dtype)
|
||||
|
||||
quantiles = [0.5, 0.2, 0.8]
|
||||
|
||||
if provider == "torch-bf16":
|
||||
ms, min_ms, max_ms = triton.testing.do_bench_cudagraph(
|
||||
lambda: torch.nn.functional.linear(a, b), quantiles=quantiles
|
||||
)
|
||||
else:
|
||||
cfg = PROVIDER_CFGS[provider]
|
||||
run_quant = build_nvfp4_runner(cfg, a, b, dtype, device)
|
||||
ms, min_ms, max_ms = triton.testing.do_bench_cudagraph(
|
||||
lambda: run_quant(), quantiles=quantiles
|
||||
)
|
||||
|
||||
to_tflops = lambda t_ms: (2 * M * N * K) * 1e-12 / (t_ms * 1e-3)
|
||||
return to_tflops(ms), to_tflops(max_ms), to_tflops(min_ms)
|
||||
|
||||
|
||||
def prepare_shapes(args):
|
||||
out = []
|
||||
for model, tp_size in itertools.product(args.models, args.tp_sizes):
|
||||
for KN, tp_dim in copy.deepcopy(WEIGHT_SHAPES[model]):
|
||||
KN[tp_dim] //= tp_size
|
||||
KN.append(model)
|
||||
out.append(KN)
|
||||
return out
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument(
|
||||
"--models",
|
||||
nargs="+",
|
||||
type=str,
|
||||
default=["meta-llama/Llama-3.1-8B-Instruct"],
|
||||
choices=list(WEIGHT_SHAPES.keys()),
|
||||
)
|
||||
parser.add_argument("--tp-sizes", nargs="+", type=int, default=[1])
|
||||
args = parser.parse_args()
|
||||
|
||||
for K, N, model in prepare_shapes(args):
|
||||
print(f"{model}, N={N} K={K}, BF16 vs NVFP4 GEMMs TFLOP/s:")
|
||||
benchmark.run(
|
||||
print_data=True,
|
||||
show_plots=True,
|
||||
save_path=f"bench_nvfp4_res_n{N}_k{K}",
|
||||
N=N,
|
||||
K=K,
|
||||
)
|
||||
|
||||
print("Benchmark finished!")
|
||||
@ -1,98 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import itertools
|
||||
from typing import Callable
|
||||
|
||||
import torch
|
||||
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.config import CompilationConfig, VllmConfig, set_current_vllm_config
|
||||
from vllm.model_executor.layers.quantization.input_quant_fp8 import QuantFP8
|
||||
from vllm.model_executor.layers.quantization.utils.quant_utils import GroupShape
|
||||
from vllm.triton_utils import triton
|
||||
|
||||
|
||||
# TODO(luka): use standalone_compile utility
|
||||
def with_dyn_arg(fn: Callable, arg_index: int, dim_index: int):
|
||||
def inner(*args):
|
||||
torch._dynamo.mark_dynamic(args[arg_index], dim_index)
|
||||
return fn(*args)
|
||||
|
||||
return inner
|
||||
|
||||
|
||||
torch._dynamo.config.recompile_limit = 8888
|
||||
compilation_config = CompilationConfig(custom_ops=["none"])
|
||||
with set_current_vllm_config(VllmConfig(compilation_config=compilation_config)):
|
||||
torch_per_token_quant_fp8 = torch.compile(
|
||||
QuantFP8(False, GroupShape.PER_TOKEN),
|
||||
fullgraph=True,
|
||||
dynamic=False, # recompile for different shapes
|
||||
)
|
||||
|
||||
# First dim is explicitly dynamic to simulate vLLM usage
|
||||
torch_per_token_quant_fp8 = with_dyn_arg(torch_per_token_quant_fp8, 0, 0)
|
||||
|
||||
|
||||
def cuda_per_token_quant_fp8(
|
||||
input: torch.Tensor,
|
||||
) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
return ops.scaled_fp8_quant(input)
|
||||
|
||||
|
||||
def calculate_diff(batch_size: int, seq_len: int):
|
||||
"""Calculate difference between Triton and CUDA implementations."""
|
||||
device = torch.device("cuda")
|
||||
x = torch.rand((batch_size * seq_len, 4096), dtype=torch.float16, device=device)
|
||||
|
||||
torch_out, torch_scale = torch_per_token_quant_fp8(x)
|
||||
cuda_out, cuda_scale = cuda_per_token_quant_fp8(x)
|
||||
|
||||
if torch.allclose(
|
||||
cuda_out.to(torch.float32), torch_out.to(torch.float32), rtol=1e-3, atol=1e-5
|
||||
) and torch.allclose(cuda_scale, torch_scale, rtol=1e-3, atol=1e-5):
|
||||
print("✅ All implementations match")
|
||||
else:
|
||||
print("❌ Implementations differ")
|
||||
|
||||
|
||||
batch_size_range = [1, 16, 32, 64, 128]
|
||||
seq_len_range = [1, 16, 64, 128, 256, 512, 1024, 2048, 4096]
|
||||
|
||||
configs = list(itertools.product(batch_size_range, seq_len_range))
|
||||
|
||||
|
||||
@triton.testing.perf_report(
|
||||
triton.testing.Benchmark(
|
||||
x_names=["batch_size", "seq_len"],
|
||||
x_vals=configs,
|
||||
line_arg="provider",
|
||||
line_vals=["torch", "cuda"],
|
||||
line_names=["Torch", "CUDA"],
|
||||
styles=[("blue", "-"), ("green", "-")],
|
||||
ylabel="us",
|
||||
plot_name="per-token-dynamic-quant-fp8-performance",
|
||||
args={},
|
||||
)
|
||||
)
|
||||
def benchmark_quantization(batch_size, seq_len, provider):
|
||||
dtype = torch.float16
|
||||
device = torch.device("cuda")
|
||||
|
||||
x = torch.randn(batch_size * seq_len, 4096, device=device, dtype=dtype)
|
||||
|
||||
quantiles = [0.5, 0.2, 0.8]
|
||||
|
||||
if provider == "torch":
|
||||
fn = lambda: torch_per_token_quant_fp8(x.clone())
|
||||
elif provider == "cuda":
|
||||
fn = lambda: cuda_per_token_quant_fp8(x.clone())
|
||||
|
||||
ms, min_ms, max_ms = triton.testing.do_bench_cudagraph(fn, quantiles=quantiles)
|
||||
|
||||
return 1000 * ms, 1000 * max_ms, 1000 * min_ms
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
calculate_diff(batch_size=4, seq_len=4096)
|
||||
benchmark_quantization.run(print_data=True)
|
||||
345
benchmarks/kernels/benchmark_aqlm.py
Normal file
345
benchmarks/kernels/benchmark_aqlm.py
Normal file
@ -0,0 +1,345 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import os
|
||||
import sys
|
||||
from typing import Optional
|
||||
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.model_executor.layers.quantization.aqlm import (
|
||||
dequantize_weight,
|
||||
generic_dequantize_gemm,
|
||||
get_int_dtype,
|
||||
optimized_dequantize_gemm,
|
||||
)
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
|
||||
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
||||
|
||||
|
||||
def torch_mult(
|
||||
# [..., in_features]
|
||||
input: torch.Tensor,
|
||||
weights: torch.Tensor,
|
||||
# [num_out_groups, 1, 1, 1]
|
||||
scales: torch.Tensor,
|
||||
) -> torch.Tensor:
|
||||
output = F.linear(input, weights)
|
||||
return output
|
||||
|
||||
|
||||
def dequant_out_scale(
|
||||
# [..., in_features]
|
||||
input: torch.Tensor,
|
||||
# [num_out_groups, num_in_groups, num_codebooks]
|
||||
codes: torch.IntTensor,
|
||||
# [num_codebooks, codebook_size, out_group_size, in_group_size]
|
||||
codebooks: torch.Tensor,
|
||||
# [num_out_groups, 1, 1, 1]
|
||||
scales: torch.Tensor,
|
||||
output_partition_sizes: torch.IntTensor,
|
||||
bias: Optional[torch.Tensor],
|
||||
) -> torch.Tensor:
|
||||
weights = ops.aqlm_dequant(codes, codebooks, output_partition_sizes)
|
||||
|
||||
if bias is None:
|
||||
output = F.linear(input, weights, bias)
|
||||
orig_shape = output.shape
|
||||
flattened_output = output.view(-1, output.size(-1))
|
||||
f_scales = scales.view(-1, scales.shape[0])
|
||||
b_scales = f_scales.expand(flattened_output.shape[0], -1)
|
||||
flattened_output *= b_scales
|
||||
return flattened_output.view(orig_shape)
|
||||
else:
|
||||
b_scales = scales.view(scales.shape[:-3] + (-1,)).expand(-1, weights.shape[1])
|
||||
weights *= b_scales
|
||||
return F.linear(input, weights, bias)
|
||||
|
||||
|
||||
def dequant_weight_scale(
|
||||
# [..., in_features]
|
||||
input: torch.Tensor,
|
||||
# [num_out_groups, num_in_groups, num_codebooks]
|
||||
codes: torch.IntTensor,
|
||||
# [num_codebooks, codebook_size, out_group_size, in_group_size]
|
||||
codebooks: torch.Tensor,
|
||||
# [num_out_groups, 1, 1, 1]
|
||||
scales: torch.Tensor,
|
||||
output_partition_sizes: torch.IntTensor,
|
||||
bias: Optional[torch.Tensor],
|
||||
) -> torch.Tensor:
|
||||
weights = ops.aqlm_dequant(codes, codebooks, output_partition_sizes)
|
||||
|
||||
b_scales = scales.view(scales.shape[:-3] + (-1,)).expand(-1, weights.shape[1])
|
||||
weights *= b_scales
|
||||
return F.linear(input, weights, bias)
|
||||
|
||||
|
||||
def dequant_no_scale(
|
||||
# [..., in_features]
|
||||
input: torch.Tensor,
|
||||
# [num_out_groups, num_in_groups, num_codebooks]
|
||||
codes: torch.IntTensor,
|
||||
# [num_codebooks, codebook_size, out_group_size, in_group_size]
|
||||
codebooks: torch.Tensor,
|
||||
# [num_out_groups, 1, 1, 1]
|
||||
scales: torch.Tensor,
|
||||
output_partition_sizes: torch.IntTensor,
|
||||
bias: Optional[torch.Tensor],
|
||||
) -> torch.Tensor:
|
||||
weights = ops.aqlm_dequant(codes, codebooks, output_partition_sizes)
|
||||
|
||||
return F.linear(input, weights, bias)
|
||||
|
||||
|
||||
# Compare the optimized 1x16 and 2x8 cuda decompression/dequant kernels against
|
||||
# the generic pytorch version.
|
||||
# Just visual comparison.
|
||||
def dequant_test(k: int, parts: torch.Tensor, nbooks: int, bits: int) -> None:
|
||||
n = int(parts.sum().item())
|
||||
|
||||
device = torch.device("cuda:0")
|
||||
|
||||
code_range = (1 << bits) // 2
|
||||
ingroups = 8
|
||||
|
||||
codes = torch.randint(
|
||||
-code_range,
|
||||
code_range,
|
||||
size=(n, k // ingroups, nbooks),
|
||||
dtype=get_int_dtype(bits),
|
||||
device=device,
|
||||
)
|
||||
|
||||
codebooks = torch.randn(
|
||||
size=(parts.shape[0] * nbooks, 1 << bits, 1, 8),
|
||||
dtype=torch.float16,
|
||||
device=device,
|
||||
)
|
||||
|
||||
count = 0
|
||||
for index in range(16):
|
||||
for i in range(8):
|
||||
for book in range(nbooks):
|
||||
codebooks[book, index, 0, i] = count * (10**book)
|
||||
count += 1
|
||||
|
||||
print("codes shape", codes.shape)
|
||||
|
||||
for i in range(16):
|
||||
for book in range(nbooks):
|
||||
codes[0, i, book] = i
|
||||
codes[0, -i, book] = i
|
||||
|
||||
weights = dequantize_weight(codes, codebooks, None)
|
||||
weights2 = ops.aqlm_dequant(codes, codebooks, parts)
|
||||
|
||||
print("weights shape:", weights.shape)
|
||||
print("weights2 shape:", weights2.shape)
|
||||
|
||||
print("weights are:", weights)
|
||||
print("weights2 are:", weights2)
|
||||
|
||||
print("first 128 weights are", weights[0, 0:128].to(torch.int32))
|
||||
print("first 128 weights2 are:", weights2[0, 0:128].to(torch.int32))
|
||||
|
||||
print("last 128 weights are", weights[0, -128:])
|
||||
print("last 128 weights2 are:", weights2[0, -128:])
|
||||
|
||||
|
||||
def main():
|
||||
parser = FlexibleArgumentParser(description="Benchmark aqlm performance.")
|
||||
|
||||
# Add arguments
|
||||
parser.add_argument(
|
||||
"--nbooks", type=int, default=1, help="Number of codebooks (default: 1)"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--bits",
|
||||
type=int,
|
||||
default=16,
|
||||
help="Number of bits per code element (default: 16)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--test",
|
||||
type=bool,
|
||||
default=False,
|
||||
help="Run the decompression/dequant tester rather than benchmarking "
|
||||
"(default: False)",
|
||||
)
|
||||
|
||||
# Parse the arguments
|
||||
args = parser.parse_args()
|
||||
|
||||
# Extract values
|
||||
nbooks = args.nbooks
|
||||
bits = args.bits
|
||||
|
||||
if args.test:
|
||||
dequant_test(4096, torch.tensor((4096,)), nbooks, bits)
|
||||
return
|
||||
|
||||
# Otherwise, benchmark.
|
||||
methods = [
|
||||
ops.aqlm_gemm,
|
||||
dequant_out_scale,
|
||||
generic_dequantize_gemm,
|
||||
optimized_dequantize_gemm,
|
||||
dequant_weight_scale,
|
||||
torch_mult,
|
||||
dequant_no_scale,
|
||||
]
|
||||
|
||||
filename = f"./aqlm_benchmark_{nbooks}x{bits}.csv"
|
||||
print(f"writing benchmarks to file {filename}")
|
||||
with open(filename, "w") as f:
|
||||
sys.stdout = f
|
||||
|
||||
print("m | k | n | n parts", end="")
|
||||
for method in methods:
|
||||
print(f" | {method.__name__.replace('_', ' ')} (µs)", end="")
|
||||
print("")
|
||||
|
||||
# These are reasonable prefill sizes.
|
||||
ksandpartions = (
|
||||
(4096, (4096, 4096, 4096)),
|
||||
(4096, (4096,)),
|
||||
(4096, (11008, 11008)),
|
||||
(11008, (4096,)),
|
||||
)
|
||||
|
||||
# reasonable ranges for m.
|
||||
for m in [
|
||||
1,
|
||||
2,
|
||||
4,
|
||||
8,
|
||||
10,
|
||||
12,
|
||||
14,
|
||||
16,
|
||||
24,
|
||||
32,
|
||||
48,
|
||||
52,
|
||||
56,
|
||||
64,
|
||||
96,
|
||||
112,
|
||||
128,
|
||||
256,
|
||||
512,
|
||||
1024,
|
||||
1536,
|
||||
2048,
|
||||
3072,
|
||||
4096,
|
||||
]:
|
||||
print(f"{m}", file=sys.__stdout__)
|
||||
for ksp in ksandpartions:
|
||||
run_grid(m, ksp[0], torch.tensor(ksp[1]), nbooks, bits, methods)
|
||||
|
||||
sys.stdout = sys.__stdout__
|
||||
|
||||
|
||||
def run_grid(m: int, k: int, parts: torch.Tensor, nbooks: int, bits: int, methods):
|
||||
# I didn't see visible improvements from increasing these, but feel free :)
|
||||
num_warmup_trials = 1
|
||||
num_trials = 1
|
||||
|
||||
num_calls = 100
|
||||
|
||||
# warmup.
|
||||
for method in methods:
|
||||
for _ in range(num_warmup_trials):
|
||||
run_timing(
|
||||
num_calls=num_calls,
|
||||
m=m,
|
||||
k=k,
|
||||
parts=parts,
|
||||
nbooks=nbooks,
|
||||
bits=bits,
|
||||
method=method,
|
||||
)
|
||||
|
||||
n = parts.sum().item()
|
||||
print(f"{m} | {k} | {n} | {parts.tolist()}", end="")
|
||||
|
||||
for method in methods:
|
||||
best_time_us = 1e20
|
||||
for _ in range(num_trials):
|
||||
kernel_dur_ms = run_timing(
|
||||
num_calls=num_calls,
|
||||
m=m,
|
||||
k=k,
|
||||
parts=parts,
|
||||
nbooks=nbooks,
|
||||
bits=bits,
|
||||
method=method,
|
||||
)
|
||||
|
||||
kernel_dur_us = 1000 * kernel_dur_ms
|
||||
|
||||
if kernel_dur_us < best_time_us:
|
||||
best_time_us = kernel_dur_us
|
||||
|
||||
print(f" | {kernel_dur_us:.0f}", end="")
|
||||
|
||||
print("")
|
||||
|
||||
|
||||
def run_timing(
|
||||
num_calls: int, m: int, k: int, parts: torch.Tensor, nbooks: int, bits: int, method
|
||||
) -> float:
|
||||
n = int(parts.sum().item())
|
||||
|
||||
device = torch.device("cuda:0")
|
||||
|
||||
input = torch.randn((1, m, k), dtype=torch.float16, device=device)
|
||||
|
||||
code_range = (1 << bits) // 2
|
||||
ingroups = 8
|
||||
|
||||
codes = torch.randint(
|
||||
-code_range,
|
||||
code_range,
|
||||
size=(n, k // ingroups, nbooks),
|
||||
dtype=get_int_dtype(bits),
|
||||
device=device,
|
||||
)
|
||||
|
||||
codebooks = torch.randn(
|
||||
size=(parts.shape[0] * nbooks, 1 << bits, 1, 8),
|
||||
dtype=torch.float16,
|
||||
device=device,
|
||||
)
|
||||
|
||||
scales = torch.randn(size=(n, 1, 1, 1), dtype=torch.float16, device=device)
|
||||
|
||||
# for comparison to just a pytorch mult.
|
||||
weights = torch.randn((n, k), dtype=torch.float16, device=device)
|
||||
|
||||
start_event = torch.cuda.Event(enable_timing=True)
|
||||
end_event = torch.cuda.Event(enable_timing=True)
|
||||
|
||||
start_event.record()
|
||||
|
||||
if method is torch_mult:
|
||||
for i in range(num_calls):
|
||||
torch_mult(input, weights, scales)
|
||||
else:
|
||||
for i in range(num_calls):
|
||||
method(input, codes, codebooks, scales, parts, None)
|
||||
|
||||
end_event.record()
|
||||
end_event.synchronize()
|
||||
|
||||
dur_ms = start_event.elapsed_time(end_event) / num_calls
|
||||
return dur_ms
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
sys.exit(main())
|
||||
@ -3,8 +3,6 @@
|
||||
# Copyright (c) Microsoft Corporation.
|
||||
# Licensed under the MIT License.
|
||||
|
||||
from packaging import version
|
||||
|
||||
from vllm.model_executor.layers.quantization.utils.bitblas_utils import (
|
||||
MINIMUM_BITBLAS_VERSION,
|
||||
)
|
||||
@ -12,7 +10,7 @@ from vllm.model_executor.layers.quantization.utils.bitblas_utils import (
|
||||
try:
|
||||
import bitblas
|
||||
|
||||
if version.parse(bitblas.__version__) < version.parse(MINIMUM_BITBLAS_VERSION):
|
||||
if bitblas.__version__ < MINIMUM_BITBLAS_VERSION:
|
||||
raise ImportError(
|
||||
"bitblas version is wrong. Please "
|
||||
f"install bitblas>={MINIMUM_BITBLAS_VERSION}"
|
||||
|
||||
@ -80,11 +80,6 @@ def bench_run(
|
||||
a, score, topk, renormalize=False
|
||||
)
|
||||
|
||||
ab_strides1 = torch.full((num_experts,), k, device="cuda", dtype=torch.int64)
|
||||
ab_strides2 = torch.full((num_experts,), n, device="cuda", dtype=torch.int64)
|
||||
c_strides1 = torch.full((num_experts,), 2 * n, device="cuda", dtype=torch.int64)
|
||||
c_strides2 = torch.full((num_experts,), k, device="cuda", dtype=torch.int64)
|
||||
|
||||
def run_triton_moe(
|
||||
a: torch.Tensor,
|
||||
w1: torch.Tensor,
|
||||
@ -116,10 +111,6 @@ def bench_run(
|
||||
w2: torch.Tensor,
|
||||
w1_scale: torch.Tensor,
|
||||
w2_scale: torch.Tensor,
|
||||
ab_strides1: torch.Tensor,
|
||||
ab_strides2: torch.Tensor,
|
||||
c_strides1: torch.Tensor,
|
||||
c_strides2: torch.Tensor,
|
||||
topk_weights: torch.Tensor,
|
||||
topk_ids: torch.Tensor,
|
||||
per_act_token: bool,
|
||||
@ -134,10 +125,6 @@ def bench_run(
|
||||
topk_ids,
|
||||
w1_scale,
|
||||
w2_scale,
|
||||
ab_strides1,
|
||||
ab_strides2,
|
||||
c_strides1,
|
||||
c_strides2,
|
||||
per_act_token,
|
||||
a1_scale=None,
|
||||
)
|
||||
@ -149,10 +136,6 @@ def bench_run(
|
||||
w2_q: torch.Tensor,
|
||||
w1_scale: torch.Tensor,
|
||||
w2_scale: torch.Tensor,
|
||||
ab_strides1: torch.Tensor,
|
||||
ab_strides2: torch.Tensor,
|
||||
c_strides1: torch.Tensor,
|
||||
c_strides2: torch.Tensor,
|
||||
topk_weights: torch.Tensor,
|
||||
topk_ids: torch.Tensor,
|
||||
):
|
||||
@ -167,10 +150,6 @@ def bench_run(
|
||||
topk_ids,
|
||||
w1_scale,
|
||||
w2_scale,
|
||||
ab_strides1,
|
||||
ab_strides2,
|
||||
c_strides1,
|
||||
c_strides2,
|
||||
per_act_token,
|
||||
a1_scale=None,
|
||||
)
|
||||
@ -215,10 +194,6 @@ def bench_run(
|
||||
w2_q,
|
||||
w1_scale,
|
||||
w2_scale,
|
||||
ab_strides1,
|
||||
ab_strides2,
|
||||
c_strides1,
|
||||
c_strides2,
|
||||
topk_weights,
|
||||
topk_ids,
|
||||
)
|
||||
@ -256,10 +231,6 @@ def bench_run(
|
||||
"w1_scale": w1_scale,
|
||||
"w2_scale": w2_scale,
|
||||
"per_act_token": per_act_token,
|
||||
"ab_strides1": ab_strides1,
|
||||
"ab_strides2": ab_strides2,
|
||||
"c_strides1": c_strides1,
|
||||
"c_strides2": c_strides2,
|
||||
# cuda graph params
|
||||
"cutlass_graph": cutlass_graph,
|
||||
"triton_graph": triton_graph,
|
||||
@ -318,10 +289,6 @@ def bench_run(
|
||||
w2_q,
|
||||
w1_scale,
|
||||
w2_scale,
|
||||
ab_strides1,
|
||||
ab_strides2,
|
||||
c_strides1,
|
||||
c_strides2,
|
||||
topk_weights,
|
||||
topk_ids,
|
||||
per_act_token,
|
||||
@ -330,7 +297,7 @@ def bench_run(
|
||||
|
||||
results.append(
|
||||
benchmark.Timer(
|
||||
stmt="run_cutlass_moe(a, a_scale, w1_q, w2_q, w1_scale, w2_scale, ab_strides1, ab_strides2, c_strides1, c_strides2, topk_weights, topk_ids, per_act_token, num_runs)", # noqa: E501
|
||||
stmt="run_cutlass_moe(a, a_scale, w1_q, w2_q, w1_scale, w2_scale, topk_weights, topk_ids, per_act_token, num_runs)", # noqa: E501
|
||||
globals=globals,
|
||||
label=label,
|
||||
sub_label=sub_label,
|
||||
|
||||
@ -637,7 +637,7 @@ def bench_optype(
|
||||
# Clear LoRA optimization hash-maps.
|
||||
_LORA_A_PTR_DICT.clear()
|
||||
_LORA_B_PTR_DICT.clear()
|
||||
# Run bench function so that _LORA_A_PTR_DICT and _LORA_B_PTR_DICT are set up
|
||||
# Run bench function so that _LORA_A_PTR_DICT and _LORA_B_PTR_DICT are setup
|
||||
for kwargs in kwargs_list:
|
||||
op_type.bench_fn()(**kwargs)
|
||||
torch.cuda.synchronize()
|
||||
|
||||
@ -236,7 +236,6 @@ def marlin_create_bench_fn(bt: BenchmarkTensors) -> Callable:
|
||||
a=bt.a,
|
||||
c=None,
|
||||
b_q_weight=w_q,
|
||||
b_bias=None,
|
||||
b_scales=w_s,
|
||||
global_scale=None,
|
||||
b_zeros=w_zp,
|
||||
@ -253,7 +252,28 @@ def marlin_create_bench_fn(bt: BenchmarkTensors) -> Callable:
|
||||
else:
|
||||
assert bt.a.dtype == torch.int8
|
||||
assert bt.wtype == scalar_types.uint4b8
|
||||
raise NotImplementedError("QQQ is not supported anymore")
|
||||
|
||||
if bt.w_ch_s is not None:
|
||||
s_ch = bt.w_ch_s.to(torch.float32)
|
||||
else:
|
||||
s_ch = torch.ones(bt.w_ref.shape[1], dtype=torch.float32, device=device)
|
||||
|
||||
if bt.w_tok_s is not None:
|
||||
s_tok = bt.w_tok_s.to(torch.float32)
|
||||
else:
|
||||
s_tok = torch.ones(bt.a.shape[0], dtype=torch.float32, device=device)
|
||||
|
||||
fn = lambda: ops.marlin_qqq_gemm(
|
||||
a=bt.a,
|
||||
b_q_weight=w_q,
|
||||
s_group=w_s,
|
||||
s_tok=s_tok,
|
||||
s_ch=s_ch,
|
||||
workspace=workspace.scratch,
|
||||
size_m=bt.a.shape[0],
|
||||
size_n=bt.w_ref.shape[1],
|
||||
size_k=bt.w_ref.shape[0],
|
||||
)
|
||||
|
||||
return fn
|
||||
|
||||
@ -284,25 +304,6 @@ def machete_create_bench_fn(
|
||||
)
|
||||
|
||||
|
||||
def cutlass_w4a8_create_bench_fn(
|
||||
bt: BenchmarkTensors, out_type=torch.dtype, schedule=None
|
||||
) -> Callable:
|
||||
w_q = bt.w_q.t().contiguous().t() # make col major
|
||||
w_q = ops.cutlass_encode_and_reorder_int4b(w_q)
|
||||
# expects fp8 scales
|
||||
w_s = ops.cutlass_pack_scale_fp8(bt.w_g_s.to(torch.float8_e4m3fn))
|
||||
|
||||
return lambda: ops.cutlass_w4a8_mm(
|
||||
a=bt.a,
|
||||
b_q=w_q,
|
||||
b_group_scales=w_s,
|
||||
b_group_size=bt.group_size,
|
||||
b_channel_scales=bt.w_ch_s,
|
||||
a_token_scales=bt.w_tok_s,
|
||||
maybe_schedule=schedule,
|
||||
)
|
||||
|
||||
|
||||
# impl
|
||||
|
||||
# bench
|
||||
@ -404,20 +405,6 @@ def bench(
|
||||
)
|
||||
)
|
||||
|
||||
# cutlass w4a8
|
||||
if types.act_type == torch.float8_e4m3fn and group_size == 128:
|
||||
timers.append(
|
||||
bench_fns(
|
||||
label,
|
||||
sub_label,
|
||||
f"cutlass w4a8 ({name_type_string})",
|
||||
[
|
||||
cutlass_w4a8_create_bench_fn(bt, out_type=types.output_type)
|
||||
for bt in benchmark_tensors
|
||||
],
|
||||
)
|
||||
)
|
||||
|
||||
if sweep_schedules:
|
||||
global _SWEEP_SCHEDULES_RESULTS
|
||||
|
||||
|
||||
@ -3,7 +3,6 @@
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import os
|
||||
import time
|
||||
from contextlib import nullcontext
|
||||
from datetime import datetime
|
||||
@ -23,13 +22,6 @@ from vllm.utils import FlexibleArgumentParser
|
||||
FP8_DTYPE = current_platform.fp8_dtype()
|
||||
|
||||
|
||||
def ensure_divisibility(numerator, denominator, text):
|
||||
"""Ensure that numerator is divisible by the denominator."""
|
||||
assert numerator % denominator == 0, "{} {} is not divisible by tp {}.".format(
|
||||
text, numerator, denominator
|
||||
)
|
||||
|
||||
|
||||
class BenchmarkConfig(TypedDict):
|
||||
BLOCK_SIZE_M: int
|
||||
BLOCK_SIZE_N: int
|
||||
@ -94,9 +86,6 @@ def benchmark_config(
|
||||
(num_experts, 2 * shard_intermediate_size), dtype=torch.float32
|
||||
)
|
||||
w2_scale = torch.randn((hidden_size, num_experts), dtype=torch.float32)
|
||||
if use_deep_gemm:
|
||||
# we use the default block shape for deepgemm
|
||||
block_quant_shape = [128, 128]
|
||||
if use_fp8_w8a8:
|
||||
if block_quant_shape:
|
||||
block_n, block_k = block_quant_shape[0], block_quant_shape[1]
|
||||
@ -419,10 +408,8 @@ class BenchmarkWorker:
|
||||
)
|
||||
# NOTE(woosuk): The current naming convention uses w2.shape[2], which
|
||||
# is the intermediate size after silu_and_mul.
|
||||
block_n = block_quant_shape[0] if block_quant_shape else None
|
||||
block_k = block_quant_shape[1] if block_quant_shape else None
|
||||
op_config = get_moe_configs(
|
||||
num_experts, shard_intermediate_size // 2, dtype_str, block_n, block_k
|
||||
num_experts, shard_intermediate_size // 2, dtype_str
|
||||
)
|
||||
if op_config is None:
|
||||
config = get_default_config(
|
||||
@ -432,7 +419,7 @@ class BenchmarkWorker:
|
||||
hidden_size,
|
||||
topk,
|
||||
dtype_str,
|
||||
block_quant_shape,
|
||||
is_marlin=False,
|
||||
)
|
||||
else:
|
||||
config = op_config[min(op_config.keys(), key=lambda x: abs(x - num_tokens))]
|
||||
@ -545,7 +532,6 @@ def save_configs(
|
||||
use_fp8_w8a8: bool,
|
||||
use_int8_w8a16: bool,
|
||||
block_quant_shape: list[int],
|
||||
save_dir: str,
|
||||
) -> None:
|
||||
dtype_str = get_config_dtype_str(
|
||||
dtype, use_int8_w8a16=use_int8_w8a16, use_fp8_w8a8=use_fp8_w8a8
|
||||
@ -556,8 +542,7 @@ def save_configs(
|
||||
filename = get_config_file_name(
|
||||
num_experts, shard_intermediate_size // 2, dtype_str, block_quant_shape
|
||||
)
|
||||
os.makedirs(save_dir, exist_ok=True)
|
||||
filename = os.path.join(save_dir, filename)
|
||||
|
||||
print(f"Writing best config to {filename}...")
|
||||
with open(filename, "w") as f:
|
||||
json.dump(configs, f, indent=4)
|
||||
@ -582,26 +567,22 @@ def main(args: argparse.Namespace):
|
||||
E = config.ffn_config.moe_num_experts
|
||||
topk = config.ffn_config.moe_top_k
|
||||
intermediate_size = config.ffn_config.ffn_hidden_size
|
||||
shard_intermediate_size = 2 * intermediate_size // args.tp_size
|
||||
elif config.architectures[0] == "JambaForCausalLM":
|
||||
E = config.num_experts
|
||||
topk = config.num_experts_per_tok
|
||||
intermediate_size = config.intermediate_size
|
||||
elif config.architectures[0] in (
|
||||
"DeepseekV3ForCausalLM",
|
||||
"DeepseekV2ForCausalLM",
|
||||
"Glm4MoeForCausalLM",
|
||||
):
|
||||
shard_intermediate_size = 2 * intermediate_size // args.tp_size
|
||||
elif config.architectures[0] in ("DeepseekV3ForCausalLM", "DeepseekV2ForCausalLM"):
|
||||
E = config.n_routed_experts
|
||||
topk = config.num_experts_per_tok
|
||||
intermediate_size = config.moe_intermediate_size
|
||||
shard_intermediate_size = 2 * intermediate_size // args.tp_size
|
||||
elif config.architectures[0] in ("Qwen2MoeForCausalLM", "Qwen3MoeForCausalLM"):
|
||||
E = config.num_experts
|
||||
topk = config.num_experts_per_tok
|
||||
intermediate_size = config.moe_intermediate_size
|
||||
elif config.architectures[0] in ("HunYuanMoEV1ForCausalLM"):
|
||||
E = config.num_experts
|
||||
topk = config.moe_topk[0]
|
||||
intermediate_size = config.moe_intermediate_size[0]
|
||||
shard_intermediate_size = 2 * intermediate_size // args.tp_size
|
||||
else:
|
||||
# Support for llama4
|
||||
config = config.get_text_config()
|
||||
@ -609,14 +590,8 @@ def main(args: argparse.Namespace):
|
||||
E = config.num_local_experts
|
||||
topk = config.num_experts_per_tok
|
||||
intermediate_size = config.intermediate_size
|
||||
enable_ep = bool(args.enable_expert_parallel)
|
||||
if enable_ep:
|
||||
ensure_divisibility(E, args.tp_size, "Number of experts")
|
||||
E = E // args.tp_size
|
||||
shard_intermediate_size = 2 * intermediate_size
|
||||
else:
|
||||
ensure_divisibility(intermediate_size, args.tp_size, "intermediate_size")
|
||||
shard_intermediate_size = 2 * intermediate_size // args.tp_size
|
||||
|
||||
hidden_size = config.hidden_size
|
||||
dtype = torch.float16 if current_platform.is_rocm() else config.torch_dtype
|
||||
use_fp8_w8a8 = args.dtype == "fp8_w8a8"
|
||||
@ -645,7 +620,7 @@ def main(args: argparse.Namespace):
|
||||
4096,
|
||||
]
|
||||
else:
|
||||
batch_sizes = args.batch_size
|
||||
batch_sizes = [args.batch_size]
|
||||
|
||||
use_deep_gemm = bool(args.use_deep_gemm)
|
||||
|
||||
@ -712,7 +687,6 @@ def main(args: argparse.Namespace):
|
||||
use_fp8_w8a8,
|
||||
use_int8_w8a16,
|
||||
block_quant_shape,
|
||||
args.save_dir,
|
||||
)
|
||||
end = time.time()
|
||||
print(f"Tuning took {end - start:.2f} seconds")
|
||||
@ -749,16 +723,12 @@ if __name__ == "__main__":
|
||||
parser.add_argument(
|
||||
"--tp-size", "-tp", "--tensor-parallel-size", type=int, default=2
|
||||
)
|
||||
parser.add_argument("--enable-expert-parallel", "-enable-ep", action="store_true")
|
||||
parser.add_argument(
|
||||
"--dtype", type=str, choices=["auto", "fp8_w8a8", "int8_w8a16"], default="auto"
|
||||
)
|
||||
parser.add_argument("--use-deep-gemm", action="store_true")
|
||||
parser.add_argument(
|
||||
"--save-dir", type=str, default="./", help="Directory to save tuned results"
|
||||
)
|
||||
parser.add_argument("--seed", type=int, default=0)
|
||||
parser.add_argument("--batch-size", type=int, nargs="+", required=False)
|
||||
parser.add_argument("--batch-size", type=int, required=False)
|
||||
parser.add_argument("--tune", action="store_true")
|
||||
parser.add_argument("--trust-remote-code", action="store_true")
|
||||
parser.add_argument("--model-prefix", type=str, required=False)
|
||||
|
||||
@ -5,8 +5,9 @@ import itertools
|
||||
|
||||
import torch
|
||||
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.model_executor.layers.fused_moe.moe_align_block_size import (
|
||||
moe_align_block_size,
|
||||
moe_align_block_size_triton,
|
||||
)
|
||||
from vllm.triton_utils import triton
|
||||
|
||||
@ -20,6 +21,62 @@ def get_topk_ids(num_tokens: int, num_experts: int, topk: int) -> torch.Tensor:
|
||||
)
|
||||
|
||||
|
||||
def check_correctness(num_tokens, num_experts=256, block_size=256, topk=8):
|
||||
"""
|
||||
Verifies vllm vs. Triton
|
||||
"""
|
||||
topk_ids = get_topk_ids(num_tokens, num_experts, topk)
|
||||
|
||||
# 1. malloc space for triton and vllm
|
||||
# malloc enough space (max_num_tokens_padded) for the sorted ids
|
||||
max_num_tokens_padded = topk_ids.numel() + num_experts * (block_size - 1)
|
||||
sorted_ids_triton = torch.empty(
|
||||
(max_num_tokens_padded,), dtype=torch.int32, device="cuda"
|
||||
)
|
||||
sorted_ids_triton.fill_(topk_ids.numel()) # fill with sentinel value
|
||||
expert_ids_triton = torch.zeros(
|
||||
(max_num_tokens_padded // block_size,), dtype=torch.int32, device="cuda"
|
||||
)
|
||||
num_tokens_post_pad_triton = torch.empty((1,), dtype=torch.int32, device="cuda")
|
||||
|
||||
sorted_ids_vllm = torch.empty_like(sorted_ids_triton)
|
||||
sorted_ids_vllm.fill_(topk_ids.numel())
|
||||
expert_ids_vllm = torch.zeros_like(expert_ids_triton)
|
||||
num_tokens_post_pad_vllm = torch.empty_like(num_tokens_post_pad_triton)
|
||||
|
||||
# 2. run implementations
|
||||
moe_align_block_size_triton(
|
||||
topk_ids,
|
||||
num_experts,
|
||||
block_size,
|
||||
sorted_ids_triton,
|
||||
expert_ids_triton,
|
||||
num_tokens_post_pad_triton,
|
||||
)
|
||||
|
||||
ops.moe_align_block_size(
|
||||
topk_ids,
|
||||
num_experts,
|
||||
block_size,
|
||||
sorted_ids_vllm,
|
||||
expert_ids_vllm,
|
||||
num_tokens_post_pad_vllm,
|
||||
)
|
||||
print(f"✅ VLLM implementation works with {num_experts} experts!")
|
||||
|
||||
# 3. compare results
|
||||
if torch.allclose(expert_ids_triton, expert_ids_vllm) and torch.allclose(
|
||||
num_tokens_post_pad_triton, num_tokens_post_pad_vllm
|
||||
):
|
||||
print("✅ Triton and VLLM implementations match.")
|
||||
else:
|
||||
print("❌ Triton and VLLM implementations DO NOT match.")
|
||||
print("Triton expert_ids:", expert_ids_triton)
|
||||
print("VLLM expert_ids:", expert_ids_vllm)
|
||||
print("Triton num_tokens_post_pad:", num_tokens_post_pad_triton)
|
||||
print("VLLM num_tokens_post_pad:", num_tokens_post_pad_vllm)
|
||||
|
||||
|
||||
# test configurations
|
||||
num_tokens_range = [1, 16, 256, 4096]
|
||||
num_experts_range = [16, 64, 224, 256, 280, 512]
|
||||
@ -32,8 +89,8 @@ configs = list(itertools.product(num_tokens_range, num_experts_range, topk_range
|
||||
x_names=["num_tokens", "num_experts", "topk"],
|
||||
x_vals=configs,
|
||||
line_arg="provider",
|
||||
line_vals=["vllm"],
|
||||
line_names=["vLLM"],
|
||||
line_vals=["vllm", "triton"], # "triton"
|
||||
line_names=["VLLM", "Triton"], # "Triton"
|
||||
plot_name="moe-align-block-size-performance",
|
||||
args={},
|
||||
)
|
||||
@ -43,11 +100,37 @@ def benchmark(num_tokens, num_experts, topk, provider):
|
||||
block_size = 256
|
||||
topk_ids = get_topk_ids(num_tokens, num_experts, topk)
|
||||
|
||||
max_num_tokens_padded = topk_ids.numel() + num_experts * (block_size - 1)
|
||||
sorted_ids = torch.empty((max_num_tokens_padded,), dtype=torch.int32, device="cuda")
|
||||
sorted_ids.fill_(topk_ids.numel())
|
||||
max_num_m_blocks = max_num_tokens_padded // block_size
|
||||
expert_ids = torch.empty((max_num_m_blocks,), dtype=torch.int32, device="cuda")
|
||||
num_tokens_post_pad = torch.empty((1,), dtype=torch.int32, device="cuda")
|
||||
|
||||
quantiles = [0.5, 0.2, 0.8]
|
||||
|
||||
if provider == "vllm":
|
||||
ms, min_ms, max_ms = triton.testing.do_bench(
|
||||
lambda: moe_align_block_size(topk_ids, block_size, num_experts),
|
||||
lambda: ops.moe_align_block_size(
|
||||
topk_ids,
|
||||
num_experts,
|
||||
block_size,
|
||||
sorted_ids.clone(),
|
||||
expert_ids.clone(),
|
||||
num_tokens_post_pad.clone(),
|
||||
),
|
||||
quantiles=quantiles,
|
||||
)
|
||||
elif provider == "triton":
|
||||
ms, min_ms, max_ms = triton.testing.do_bench(
|
||||
lambda: moe_align_block_size_triton(
|
||||
topk_ids,
|
||||
num_experts,
|
||||
block_size,
|
||||
sorted_ids.clone(),
|
||||
expert_ids.clone(),
|
||||
num_tokens_post_pad.clone(),
|
||||
),
|
||||
quantiles=quantiles,
|
||||
)
|
||||
|
||||
@ -71,4 +154,6 @@ if __name__ == "__main__":
|
||||
)
|
||||
args = parser.parse_args()
|
||||
|
||||
print("Running correctness check...")
|
||||
check_correctness(num_tokens=1024, num_experts=args.num_experts, topk=args.topk)
|
||||
benchmark.run(print_data=True, show_plots=True)
|
||||
|
||||
@ -8,13 +8,12 @@ import ray
|
||||
import torch
|
||||
from transformers import AutoConfig
|
||||
|
||||
from vllm.model_executor.layers.fused_moe.fused_moe import *
|
||||
from vllm.model_executor.layers.fused_moe.moe_permute_unpermute import (
|
||||
from vllm.model_executor.layers.fused_moe.deep_gemm_moe import (
|
||||
_moe_permute,
|
||||
_moe_unpermute_and_reduce,
|
||||
moe_permute,
|
||||
moe_unpermute,
|
||||
)
|
||||
from vllm.model_executor.layers.fused_moe.fused_moe import *
|
||||
from vllm.model_executor.layers.fused_moe.moe_permute_unpermute import *
|
||||
from vllm.model_executor.layers.fused_moe.utils import _fp8_quantize
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
@ -64,19 +63,18 @@ def benchmark_permute(
|
||||
|
||||
def run():
|
||||
if use_customized_permute:
|
||||
(
|
||||
permuted_hidden_states,
|
||||
a1q_scale,
|
||||
first_token_off,
|
||||
inv_perm_idx,
|
||||
m_indices,
|
||||
) = moe_permute(
|
||||
qhidden_states,
|
||||
a1q_scale=None,
|
||||
topk_ids=topk_ids,
|
||||
n_expert=num_experts,
|
||||
expert_map=None,
|
||||
align_block_size=align_block_size,
|
||||
(permuted_hidden_states, first_token_off, inv_perm_idx, m_indices) = (
|
||||
moe_permute(
|
||||
qhidden_states,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
token_expert_indices=token_expert_indices,
|
||||
topk=topk,
|
||||
n_expert=num_experts,
|
||||
n_local_expert=num_experts,
|
||||
expert_map=None,
|
||||
align_block_size=align_block_size,
|
||||
)
|
||||
)
|
||||
else:
|
||||
(
|
||||
@ -152,19 +150,18 @@ def benchmark_unpermute(
|
||||
|
||||
def prepare():
|
||||
if use_customized_permute:
|
||||
(
|
||||
permuted_hidden_states,
|
||||
a1q_scale,
|
||||
first_token_off,
|
||||
inv_perm_idx,
|
||||
m_indices,
|
||||
) = moe_permute(
|
||||
qhidden_states,
|
||||
a1q_scale=None,
|
||||
topk_ids=topk_ids,
|
||||
n_expert=num_experts,
|
||||
expert_map=None,
|
||||
align_block_size=align_block_size,
|
||||
(permuted_hidden_states, first_token_off, inv_perm_idx, m_indices) = (
|
||||
moe_permute(
|
||||
qhidden_states,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
token_expert_indices=token_expert_indices,
|
||||
topk=topk,
|
||||
n_expert=num_experts,
|
||||
n_local_expert=num_experts,
|
||||
expert_map=None,
|
||||
align_block_size=align_block_size,
|
||||
)
|
||||
)
|
||||
# convert to fp16/bf16 as gemm output
|
||||
return (
|
||||
@ -194,19 +191,16 @@ def benchmark_unpermute(
|
||||
|
||||
def run(input: tuple):
|
||||
if use_customized_permute:
|
||||
(
|
||||
permuted_hidden_states,
|
||||
first_token_off,
|
||||
inv_perm_idx,
|
||||
m_indices,
|
||||
) = input
|
||||
output = torch.empty_like(hidden_states)
|
||||
(permuted_hidden_states, first_token_off, inv_perm_idx, m_indices) = input
|
||||
moe_unpermute(
|
||||
output,
|
||||
permuted_hidden_states,
|
||||
topk_weights,
|
||||
topk_ids,
|
||||
inv_perm_idx,
|
||||
first_token_off,
|
||||
topk,
|
||||
num_experts,
|
||||
num_experts,
|
||||
)
|
||||
else:
|
||||
(
|
||||
@ -217,11 +211,7 @@ def benchmark_unpermute(
|
||||
inv_perm,
|
||||
) = input
|
||||
_moe_unpermute_and_reduce(
|
||||
output_hidden_states,
|
||||
permuted_hidden_states,
|
||||
inv_perm,
|
||||
topk_weights,
|
||||
True,
|
||||
output_hidden_states, permuted_hidden_states, inv_perm, topk_weights
|
||||
)
|
||||
|
||||
# JIT compilation & warmup
|
||||
@ -328,7 +318,6 @@ def main(args: argparse.Namespace):
|
||||
elif (
|
||||
config.architectures[0] == "DeepseekV3ForCausalLM"
|
||||
or config.architectures[0] == "DeepseekV2ForCausalLM"
|
||||
or config.architectures[0] == "Glm4MoeForCausalLM"
|
||||
):
|
||||
E = config.n_routed_experts
|
||||
topk = config.num_experts_per_tok
|
||||
|
||||
@ -1,328 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
# This script benchmarks the mrope kernel (mainly for Qwen2VL and Qwen2.5VL models).
|
||||
# It generates test data, runs benchmarks, and saves results to a CSV file.
|
||||
#
|
||||
# The CSV file (named with current date/time) contains these columns:
|
||||
# model_name, tp_size, num_tokens, num_heads, num_kv_heads, head_dim, max_position,
|
||||
# rope_theta, is_neox_style, rope_scaling, dtype, torch_mean, torch_median, torch_p99,
|
||||
# torch_min, torch_max, triton_mean, triton_median, triton_p99, triton_min, triton_max,
|
||||
# speedup
|
||||
#
|
||||
# == Usage Examples ==
|
||||
#
|
||||
# Single model benchmark:
|
||||
# python3 benchmark_mrope.py --model-name Qwen/Qwen2-VL-7B-Instruct --tp-size 1 \
|
||||
# --warmup-iter 10 --benchmark-iter 100 --dtype bfloat16 --seed 0 --num-tokens 1024
|
||||
#
|
||||
# All models benchmark:
|
||||
# python3 benchmark_mrope.py --model-name "" --tp-size 1 --warmup-iter 10 \
|
||||
# --benchmark-iter 100 --dtype bfloat16 --seed 0 --num-tokens 1024
|
||||
#
|
||||
# All models with different TP sizes:
|
||||
# python3 benchmark_mrope.py --model-name "" --tp-size 1 2 4 8 --warmup-iter 10 \
|
||||
# --benchmark-iter 100 --dtype bfloat16 --seed 0 --num-tokens 1024
|
||||
#
|
||||
# All models with different token counts:
|
||||
# python3 benchmark_mrope.py --model-name "" --tp-size 1 --warmup-iter 10 \
|
||||
# --benchmark-iter 100 --dtype bfloat16 --seed 0 --num-tokens 1024 4096 16384
|
||||
import csv
|
||||
import os
|
||||
import time
|
||||
from datetime import datetime
|
||||
from typing import Any
|
||||
|
||||
import numpy as np
|
||||
import torch
|
||||
|
||||
from vllm.model_executor.layers.rotary_embedding import get_rope
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.transformers_utils.config import get_config
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
|
||||
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
||||
|
||||
|
||||
def generate_test_data(
|
||||
num_tokens: int,
|
||||
num_q_heads: int,
|
||||
num_kv_heads: int,
|
||||
head_size: int,
|
||||
max_position_embeddings: int,
|
||||
dtype: torch.dtype,
|
||||
device: torch.device,
|
||||
):
|
||||
"""Generate test data for given configuration."""
|
||||
# Create 2D positions (3, num_tokens) for multimodal case
|
||||
positions = torch.randint(
|
||||
0, max_position_embeddings // 4, (3, num_tokens), device=device
|
||||
)
|
||||
|
||||
# Create query and key tensors
|
||||
query = torch.randn(num_tokens, num_q_heads * head_size, dtype=dtype, device=device)
|
||||
key = torch.randn(num_tokens, num_kv_heads * head_size, dtype=dtype, device=device)
|
||||
|
||||
return positions, query, key
|
||||
|
||||
|
||||
def calculate_stats(times: list[float]) -> dict[str, float]:
|
||||
"""Calculate statistics from a list of times."""
|
||||
times_array = np.array(times)
|
||||
return {
|
||||
"mean": np.mean(times_array),
|
||||
"median": np.median(times_array),
|
||||
"p99": np.percentile(times_array, 99),
|
||||
"min": np.min(times_array),
|
||||
"max": np.max(times_array),
|
||||
}
|
||||
|
||||
|
||||
def benchmark_mrope(
|
||||
model_name: str,
|
||||
num_tokens: int,
|
||||
head_dim: int,
|
||||
tp_size: int,
|
||||
num_heads: int,
|
||||
num_kv_heads: int,
|
||||
max_position: int = 8192,
|
||||
rope_theta: float = 10000,
|
||||
is_neox_style: bool = True,
|
||||
rope_scaling: dict[str, Any] = None,
|
||||
dtype: torch.dtype = torch.bfloat16,
|
||||
seed: int = 0,
|
||||
warmup_iter: int = 10,
|
||||
benchmark_iter: int = 100,
|
||||
csv_writer=None,
|
||||
):
|
||||
current_platform.seed_everything(seed)
|
||||
torch.set_default_device(device)
|
||||
# the parameters to compute the q k v size based on tp_size
|
||||
mrope_helper_class = get_rope(
|
||||
head_size=head_dim,
|
||||
rotary_dim=head_dim,
|
||||
max_position=max_position,
|
||||
base=rope_theta,
|
||||
is_neox_style=is_neox_style,
|
||||
rope_scaling=rope_scaling,
|
||||
dtype=dtype,
|
||||
).to(device=device)
|
||||
|
||||
print(80 * "=")
|
||||
print(
|
||||
f"Evaluating model: {model_name} "
|
||||
f"with tp_size: {tp_size} "
|
||||
f"and num_tokens: {num_tokens}, "
|
||||
f"dtype: {dtype}"
|
||||
)
|
||||
|
||||
# create q k v input tensors
|
||||
# create rotary pos emb input tensors
|
||||
positions, query, key = generate_test_data(
|
||||
num_tokens, num_heads, num_kv_heads, head_dim, max_position, dtype, device
|
||||
)
|
||||
|
||||
# Warm up
|
||||
for _ in range(warmup_iter):
|
||||
mrope_helper_class.forward_native(
|
||||
positions,
|
||||
query.clone(),
|
||||
key.clone(),
|
||||
)
|
||||
|
||||
mrope_helper_class.forward_cuda(
|
||||
positions,
|
||||
query.clone(),
|
||||
key.clone(),
|
||||
)
|
||||
|
||||
torch.cuda.synchronize()
|
||||
|
||||
# Time reference implementation
|
||||
torch_times = []
|
||||
for _ in range(benchmark_iter):
|
||||
query_clone = query.clone()
|
||||
key_clone = key.clone()
|
||||
torch.cuda.synchronize()
|
||||
start_time = time.time()
|
||||
|
||||
mrope_helper_class.forward_native(
|
||||
positions,
|
||||
query_clone,
|
||||
key_clone,
|
||||
)
|
||||
|
||||
torch.cuda.synchronize()
|
||||
torch_times.append(time.time() - start_time)
|
||||
|
||||
# Time triton kernel implementation
|
||||
triton_times = []
|
||||
for _ in range(benchmark_iter):
|
||||
query_clone = query.clone()
|
||||
key_clone = key.clone()
|
||||
torch.cuda.synchronize()
|
||||
start_time = time.time()
|
||||
mrope_helper_class.forward_cuda(
|
||||
positions,
|
||||
query_clone,
|
||||
key_clone,
|
||||
)
|
||||
torch.cuda.synchronize()
|
||||
triton_times.append(time.time() - start_time)
|
||||
|
||||
# Calculate statistics
|
||||
torch_stats = calculate_stats(torch_times)
|
||||
triton_stats = calculate_stats(triton_times)
|
||||
print(f"\nPerformance for config ({num_tokens}, {num_heads}, {num_kv_heads}):")
|
||||
|
||||
print(
|
||||
f"Torch implementation: "
|
||||
f"mean={torch_stats['mean']:.8f}s, "
|
||||
f"median={torch_stats['median']:.8f}s, "
|
||||
f"p99={torch_stats['p99']:.8f}s"
|
||||
)
|
||||
|
||||
print(
|
||||
f"Triton implementation: "
|
||||
f"mean={triton_stats['mean']:.8f}s, "
|
||||
f"median={triton_stats['median']:.8f}s, "
|
||||
f"p99={triton_stats['p99']:.8f}s"
|
||||
)
|
||||
|
||||
print(
|
||||
f"Triton Speedup over Torch: {torch_stats['mean'] / triton_stats['mean']:.8f}x"
|
||||
)
|
||||
|
||||
# Write to CSV
|
||||
if csv_writer:
|
||||
row = [
|
||||
model_name,
|
||||
tp_size,
|
||||
num_tokens,
|
||||
num_heads,
|
||||
num_kv_heads,
|
||||
head_dim,
|
||||
max_position,
|
||||
rope_theta,
|
||||
is_neox_style,
|
||||
str(rope_scaling),
|
||||
str(dtype).split(".")[-1],
|
||||
torch_stats["mean"],
|
||||
torch_stats["median"],
|
||||
torch_stats["p99"],
|
||||
torch_stats["min"],
|
||||
torch_stats["max"],
|
||||
triton_stats["mean"],
|
||||
triton_stats["median"],
|
||||
triton_stats["p99"],
|
||||
triton_stats["min"],
|
||||
triton_stats["max"],
|
||||
torch_stats["mean"] / triton_stats["mean"], # speedup
|
||||
]
|
||||
csv_writer.writerow(row)
|
||||
|
||||
return torch_stats, triton_stats
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = FlexibleArgumentParser(
|
||||
description="Benchmark the rotary embedding kernels."
|
||||
)
|
||||
parser.add_argument("--model-name", type=str, default="")
|
||||
parser.add_argument("--tp-size", type=int, default=1)
|
||||
parser.add_argument("--warmup-iter", type=int, default=10)
|
||||
parser.add_argument("--benchmark-iter", type=int, default=100)
|
||||
parser.add_argument("--dtype", type=str, choices=["bfloat16"], default="bfloat16")
|
||||
parser.add_argument("--seed", type=int, default=0)
|
||||
parser.add_argument("--num-tokens", type=int, nargs="+", required=False)
|
||||
parser.add_argument("--trust-remote-code", action="store_true")
|
||||
parser.add_argument("--output-csv", type=str, default="mrope_benchmark_results.csv")
|
||||
args = parser.parse_args()
|
||||
print(args)
|
||||
|
||||
# Create CSV file for results
|
||||
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
||||
csv_filename = f"{os.path.splitext(args.output_csv)[0]}_{timestamp}.csv"
|
||||
|
||||
with open(csv_filename, "w", newline="") as csvfile:
|
||||
csv_writer = csv.writer(csvfile)
|
||||
# Write header
|
||||
header = [
|
||||
"model_name",
|
||||
"tp_size",
|
||||
"num_tokens",
|
||||
"num_heads",
|
||||
"num_kv_heads",
|
||||
"head_dim",
|
||||
"max_position",
|
||||
"rope_theta",
|
||||
"is_neox_style",
|
||||
"rope_scaling",
|
||||
"dtype",
|
||||
"torch_mean",
|
||||
"torch_median",
|
||||
"torch_p99",
|
||||
"torch_min",
|
||||
"torch_max",
|
||||
"triton_mean",
|
||||
"triton_median",
|
||||
"triton_p99",
|
||||
"triton_min",
|
||||
"triton_max",
|
||||
"speedup",
|
||||
]
|
||||
csv_writer.writerow(header)
|
||||
|
||||
model_tp_dict = {}
|
||||
if args.model_name == "":
|
||||
model_tp_dict = {
|
||||
"Qwen/Qwen2-VL-2B-Instruct": [1],
|
||||
"Qwen/Qwen2-VL-7B-Instruct": [1],
|
||||
"Qwen/Qwen2-VL-72B-Instruct": [2, 4, 8],
|
||||
"Qwen/Qwen2.5-VL-3B-Instruct": [1, 2, 4, 8],
|
||||
"Qwen/Qwen2.5-VL-7B-Instruct": [1, 2, 4, 8],
|
||||
"Qwen/Qwen2.5-VL-72B-Instruct": [2, 4, 8],
|
||||
}
|
||||
else:
|
||||
model_tp_dict[args.model_name] = [args.tp_size]
|
||||
|
||||
if args.num_tokens is None:
|
||||
num_tokens_list = [2**i for i in range(0, 18)]
|
||||
else:
|
||||
num_tokens_list = args.num_tokens
|
||||
|
||||
for model_name, tp_list in model_tp_dict.items():
|
||||
config = get_config(model_name, trust_remote_code=args.trust_remote_code)
|
||||
for tp_size in tp_list:
|
||||
# get the model config
|
||||
total_num_kv_heads = config.num_key_value_heads
|
||||
total_num_heads = config.num_attention_heads
|
||||
num_heads = total_num_heads // tp_size
|
||||
num_kv_heads = max(1, total_num_kv_heads // tp_size)
|
||||
head_dim = config.hidden_size // total_num_heads
|
||||
q_size = num_heads * head_dim
|
||||
kv_size = num_kv_heads * head_dim
|
||||
is_neox_style = True
|
||||
rope_theta = config.rope_theta
|
||||
max_position = config.max_position_embeddings
|
||||
|
||||
for num_tokens in num_tokens_list:
|
||||
benchmark_mrope(
|
||||
model_name=model_name,
|
||||
num_tokens=num_tokens,
|
||||
head_dim=head_dim,
|
||||
tp_size=tp_size,
|
||||
num_heads=num_heads,
|
||||
num_kv_heads=num_kv_heads,
|
||||
max_position=max_position,
|
||||
rope_theta=rope_theta,
|
||||
is_neox_style=is_neox_style,
|
||||
rope_scaling=config.rope_scaling,
|
||||
dtype=getattr(torch, args.dtype),
|
||||
seed=args.seed,
|
||||
warmup_iter=args.warmup_iter,
|
||||
benchmark_iter=args.benchmark_iter,
|
||||
csv_writer=csv_writer,
|
||||
)
|
||||
|
||||
print(f"Benchmark results saved to {csv_filename}")
|
||||
@ -1,159 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import argparse
|
||||
import math
|
||||
from contextlib import contextmanager
|
||||
from typing import Callable
|
||||
from unittest.mock import patch
|
||||
|
||||
import torch
|
||||
|
||||
from vllm.model_executor.layers.quantization.utils import fp8_utils, int8_utils
|
||||
from vllm.platforms import current_platform
|
||||
|
||||
|
||||
@contextmanager
|
||||
def _triton_mode():
|
||||
"""Temporarily force the Triton fallback path"""
|
||||
with patch("vllm.platforms.current_platform.is_cuda", return_value=False):
|
||||
yield
|
||||
|
||||
|
||||
def _time_cuda(
|
||||
fn: Callable[[], tuple[torch.Tensor, torch.Tensor]],
|
||||
warmup_iters: int,
|
||||
bench_iters: int,
|
||||
) -> float:
|
||||
# warmup
|
||||
for _ in range(warmup_iters):
|
||||
fn()
|
||||
torch.cuda.synchronize()
|
||||
|
||||
start = torch.cuda.Event(enable_timing=True)
|
||||
end = torch.cuda.Event(enable_timing=True)
|
||||
|
||||
start.record()
|
||||
for _ in range(bench_iters):
|
||||
fn()
|
||||
end.record()
|
||||
torch.cuda.synchronize()
|
||||
|
||||
return start.elapsed_time(end) / bench_iters # ms/iter
|
||||
|
||||
|
||||
def _run_single(
|
||||
shape: tuple[int, int],
|
||||
group_size: int,
|
||||
dtype: str,
|
||||
*,
|
||||
column_major: bool = False,
|
||||
scale_ue8m0: bool = False,
|
||||
warmup_iters: int,
|
||||
bench_iters: int,
|
||||
) -> None:
|
||||
num_tokens, hidden_dim = shape
|
||||
|
||||
device = torch.device("cuda")
|
||||
torch.manual_seed(42)
|
||||
x = torch.randn(num_tokens, hidden_dim, device=device, dtype=torch.bfloat16) * 8
|
||||
|
||||
if dtype == "fp8":
|
||||
|
||||
def cuda_impl():
|
||||
return fp8_utils.per_token_group_quant_fp8(
|
||||
x,
|
||||
group_size,
|
||||
column_major_scales=column_major,
|
||||
use_ue8m0=scale_ue8m0,
|
||||
)
|
||||
|
||||
def triton_impl():
|
||||
with _triton_mode():
|
||||
return fp8_utils.per_token_group_quant_fp8(
|
||||
x,
|
||||
group_size,
|
||||
column_major_scales=column_major,
|
||||
use_ue8m0=scale_ue8m0,
|
||||
)
|
||||
elif dtype == "int8":
|
||||
|
||||
def cuda_impl():
|
||||
return int8_utils.per_token_group_quant_int8(x, group_size)
|
||||
|
||||
def triton_impl():
|
||||
with _triton_mode():
|
||||
return int8_utils.per_token_group_quant_int8(x, group_size)
|
||||
else:
|
||||
raise ValueError("dtype must be 'fp8' or 'int8'")
|
||||
|
||||
cuda_ms = _time_cuda(cuda_impl, warmup_iters, bench_iters)
|
||||
triton_ms = _time_cuda(triton_impl, warmup_iters, bench_iters)
|
||||
|
||||
speedup = triton_ms / cuda_ms if cuda_ms else math.inf
|
||||
|
||||
cfg_desc = (
|
||||
f"shape={shape} gs={group_size:<3} col_major={column_major:<5} "
|
||||
f"ue8m0={scale_ue8m0:<5} dtype={dtype}"
|
||||
)
|
||||
print(
|
||||
f"{cfg_desc:55} | CUDA {cuda_ms:7.3f} ms | Triton {triton_ms:7.3f} ms | "
|
||||
f"speed-up ×{speedup:5.2f}"
|
||||
)
|
||||
|
||||
|
||||
def parse_args():
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("--warmup-iters", type=int, default=10)
|
||||
parser.add_argument("--bench-iters", type=int, default=100)
|
||||
parser.add_argument("--dtype", choices=["fp8", "int8", "both"], default="both")
|
||||
return parser.parse_args()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
if not current_platform.is_cuda():
|
||||
raise RuntimeError("CUDA device is required to run this benchmark.")
|
||||
|
||||
args = parse_args()
|
||||
warmup_iters, bench_iters = args.warmup_iters, args.bench_iters
|
||||
|
||||
shapes = [(32, 128), (64, 256), (16, 512)]
|
||||
group_sizes = [64, 128]
|
||||
|
||||
dtypes = ["fp8", "int8"] if args.dtype == "both" else [args.dtype]
|
||||
|
||||
header = (
|
||||
"Configuration".ljust(55)
|
||||
+ " | "
|
||||
+ "CUDA (ms)".center(12)
|
||||
+ " | "
|
||||
+ "Triton (ms)".center(13)
|
||||
+ " | "
|
||||
+ "Speed-up"
|
||||
)
|
||||
print(header)
|
||||
print("-" * len(header))
|
||||
|
||||
for dtype in dtypes:
|
||||
for shape in shapes:
|
||||
for gs in group_sizes:
|
||||
if dtype == "fp8":
|
||||
for col_major in (False, True):
|
||||
for ue8m0 in (False, True):
|
||||
_run_single(
|
||||
shape,
|
||||
gs,
|
||||
dtype,
|
||||
column_major=col_major,
|
||||
scale_ue8m0=ue8m0,
|
||||
warmup_iters=warmup_iters,
|
||||
bench_iters=bench_iters,
|
||||
)
|
||||
else: # INT8 has no col-major / ue8m0 switches
|
||||
_run_single(
|
||||
shape,
|
||||
gs,
|
||||
dtype,
|
||||
warmup_iters=warmup_iters,
|
||||
bench_iters=bench_iters,
|
||||
)
|
||||
@ -1,156 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
from __future__ import annotations
|
||||
|
||||
import random
|
||||
import time
|
||||
|
||||
import torch
|
||||
from tabulate import tabulate
|
||||
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.logger import init_logger
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import (
|
||||
STR_DTYPE_TO_TORCH_DTYPE,
|
||||
FlexibleArgumentParser,
|
||||
create_kv_caches_with_random_flash,
|
||||
)
|
||||
|
||||
logger = init_logger(__name__)
|
||||
|
||||
|
||||
@torch.inference_mode()
|
||||
def run_benchmark(
|
||||
num_tokens: int,
|
||||
num_heads: int,
|
||||
head_size: int,
|
||||
block_size: int,
|
||||
num_blocks: int,
|
||||
dtype: torch.dtype,
|
||||
kv_cache_dtype: str,
|
||||
kv_cache_layout: str,
|
||||
num_iters: int,
|
||||
device: str = "cuda",
|
||||
) -> float:
|
||||
"""Return latency (seconds) for given num_tokens."""
|
||||
|
||||
if kv_cache_dtype == "fp8" and head_size % 16:
|
||||
raise ValueError("fp8 kv-cache requires head_size to be a multiple of 16.")
|
||||
|
||||
current_platform.seed_everything(42)
|
||||
torch.set_default_device(device)
|
||||
|
||||
# create random key / value tensors [T, H, D].
|
||||
key = torch.randn(num_tokens, num_heads, head_size, dtype=dtype, device=device)
|
||||
value = torch.randn_like(key)
|
||||
|
||||
# prepare the slot mapping.
|
||||
# each token is assigned a unique slot in the KV-cache.
|
||||
num_slots = block_size * num_blocks
|
||||
if num_tokens > num_slots:
|
||||
raise ValueError("num_tokens cannot exceed the total number of cache slots")
|
||||
slot_mapping_lst = random.sample(range(num_slots), num_tokens)
|
||||
slot_mapping = torch.tensor(slot_mapping_lst, dtype=torch.long, device=device)
|
||||
|
||||
key_caches, value_caches = create_kv_caches_with_random_flash(
|
||||
num_blocks,
|
||||
block_size,
|
||||
1, # num_layers
|
||||
num_heads,
|
||||
head_size,
|
||||
kv_cache_dtype,
|
||||
dtype,
|
||||
device=device,
|
||||
cache_layout=kv_cache_layout,
|
||||
)
|
||||
key_cache, value_cache = key_caches[0], value_caches[0]
|
||||
|
||||
# compute per-kernel scaling factors for fp8 conversion (if used).
|
||||
k_scale = (key.amax() / 64.0).to(torch.float32)
|
||||
v_scale = (value.amax() / 64.0).to(torch.float32)
|
||||
|
||||
def run_cuda_benchmark(n_iters: int) -> float:
|
||||
nonlocal key, value, key_cache, value_cache, slot_mapping
|
||||
torch.cuda.synchronize()
|
||||
start = time.perf_counter()
|
||||
for _ in range(n_iters):
|
||||
ops.reshape_and_cache_flash(
|
||||
key,
|
||||
value,
|
||||
key_cache,
|
||||
value_cache,
|
||||
slot_mapping,
|
||||
kv_cache_dtype,
|
||||
k_scale,
|
||||
v_scale,
|
||||
)
|
||||
torch.cuda.synchronize()
|
||||
end = time.perf_counter()
|
||||
return (end - start) / n_iters
|
||||
|
||||
# warm-up
|
||||
run_cuda_benchmark(3)
|
||||
|
||||
lat = run_cuda_benchmark(num_iters)
|
||||
|
||||
# free tensors to mitigate OOM when sweeping
|
||||
del key, value, key_cache, value_cache, slot_mapping
|
||||
torch.cuda.empty_cache()
|
||||
|
||||
return lat
|
||||
|
||||
|
||||
def main(args):
|
||||
rows = []
|
||||
for layout in ["NHD", "HND"]:
|
||||
for exp in range(1, 17):
|
||||
n_tok = 2**exp
|
||||
lat = run_benchmark(
|
||||
num_tokens=n_tok,
|
||||
num_heads=args.num_heads,
|
||||
head_size=args.head_size,
|
||||
block_size=args.block_size,
|
||||
num_blocks=args.num_blocks,
|
||||
dtype=STR_DTYPE_TO_TORCH_DTYPE[args.dtype],
|
||||
kv_cache_dtype=args.kv_cache_dtype,
|
||||
kv_cache_layout=layout,
|
||||
num_iters=args.iters,
|
||||
device="cuda",
|
||||
)
|
||||
rows.append([n_tok, layout, f"{lat * 1e6:.3f}"])
|
||||
|
||||
print(tabulate(rows, headers=["num_tokens", "layout", "latency (µs)"]))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = FlexibleArgumentParser()
|
||||
|
||||
parser.add_argument("--num-heads", type=int, default=128)
|
||||
parser.add_argument(
|
||||
"--head-size",
|
||||
type=int,
|
||||
choices=[64, 80, 96, 112, 120, 128, 192, 256],
|
||||
default=128,
|
||||
)
|
||||
parser.add_argument("--block-size", type=int, choices=[16, 32], default=16)
|
||||
parser.add_argument("--num-blocks", type=int, default=128 * 512)
|
||||
|
||||
parser.add_argument(
|
||||
"--dtype",
|
||||
type=str,
|
||||
choices=["half", "bfloat16", "float"],
|
||||
default="bfloat16",
|
||||
)
|
||||
|
||||
parser.add_argument(
|
||||
"--kv-cache-dtype",
|
||||
type=str,
|
||||
choices=["auto", "fp8"],
|
||||
default="auto",
|
||||
)
|
||||
|
||||
parser.add_argument("--iters", type=int, default=100)
|
||||
args = parser.parse_args()
|
||||
|
||||
main(args)
|
||||
@ -1,77 +0,0 @@
|
||||
#!/usr/bin/env python3
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import time
|
||||
|
||||
import torch
|
||||
|
||||
from vllm.model_executor.layers.fused_moe.batched_deep_gemm_moe import (
|
||||
silu_mul_fp8_quant_deep_gemm,
|
||||
)
|
||||
from vllm.platforms import current_platform
|
||||
|
||||
|
||||
def benchmark(E, T, H, G=128, runs=50):
|
||||
current_platform.seed_everything(42)
|
||||
y = torch.randn((E, T, 2 * H), dtype=torch.bfloat16, device="cuda")
|
||||
tokens_per_expert = torch.randint(
|
||||
T // 2, T, size=(E,), dtype=torch.int32, device="cuda"
|
||||
)
|
||||
|
||||
# Warmup
|
||||
for _ in range(10):
|
||||
silu_mul_fp8_quant_deep_gemm(y, tokens_per_expert, group_size=G)
|
||||
torch.cuda.synchronize()
|
||||
|
||||
# Benchmark
|
||||
torch.cuda.synchronize()
|
||||
start = time.perf_counter()
|
||||
for _ in range(runs):
|
||||
silu_mul_fp8_quant_deep_gemm(y, tokens_per_expert, group_size=G)
|
||||
torch.cuda.synchronize()
|
||||
|
||||
avg_time = (time.perf_counter() - start) / runs * 1000
|
||||
|
||||
# Calculate actual work done (only count valid tokens)
|
||||
actual_tokens = tokens_per_expert.sum().item()
|
||||
actual_elements = actual_tokens * H
|
||||
|
||||
# GFLOPS: operations per element = exp + 3 muls + 1 div + quantization ops ≈ 8 ops
|
||||
ops_per_element = 8
|
||||
total_ops = actual_elements * ops_per_element
|
||||
gflops = total_ops / (avg_time / 1000) / 1e9
|
||||
|
||||
# Memory bandwidth: bfloat16 inputs (2 bytes), fp8 output (1 byte), scales (4 bytes)
|
||||
input_bytes = actual_tokens * 2 * H * 2 # 2*H bfloat16 inputs
|
||||
output_bytes = actual_tokens * H * 1 # H fp8 outputs
|
||||
scale_bytes = actual_tokens * (H // G) * 4 # scales in float32
|
||||
total_bytes = input_bytes + output_bytes + scale_bytes
|
||||
memory_bw = total_bytes / (avg_time / 1000) / 1e9
|
||||
|
||||
return avg_time, gflops, memory_bw
|
||||
|
||||
|
||||
configs = [
|
||||
(8, 32, 1024),
|
||||
(16, 64, 2048),
|
||||
(32, 128, 4096),
|
||||
# DeepSeekV3 Configs
|
||||
(256, 16, 7168),
|
||||
(256, 32, 7168),
|
||||
(256, 64, 7168),
|
||||
(256, 128, 7168),
|
||||
(256, 256, 7168),
|
||||
(256, 512, 7168),
|
||||
(256, 1024, 7168),
|
||||
]
|
||||
|
||||
print(f"GPU: {torch.cuda.get_device_name()}")
|
||||
print(f"{'Config':<20} {'Time(ms)':<10} {'GFLOPS':<10} {'GB/s':<10}")
|
||||
print("-" * 50)
|
||||
|
||||
for E, T, H in configs:
|
||||
try:
|
||||
time_ms, gflops, gbps = benchmark(E, T, H)
|
||||
print(f"E={E:3d},T={T:4d},H={H:4d} {time_ms:8.3f} {gflops:8.1f} {gbps:8.1f}")
|
||||
except Exception:
|
||||
print(f"E={E:3d},T={T:4d},H={H:4d} FAILED")
|
||||
@ -1,292 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import csv
|
||||
import os
|
||||
from datetime import datetime
|
||||
from typing import Optional
|
||||
|
||||
import flashinfer
|
||||
import torch
|
||||
|
||||
from vllm.utils import round_up
|
||||
|
||||
FLOAT32_BYTES = torch.finfo(torch.float).bits // 8
|
||||
FP8_DTYPE = torch.float8_e4m3fn
|
||||
FP4_DTYPE = torch.uint8
|
||||
|
||||
|
||||
def to_float8(x, dtype=torch.float8_e4m3fn):
|
||||
finfo = torch.finfo(dtype)
|
||||
min_val, max_val = x.aminmax()
|
||||
amax = torch.maximum(min_val.abs(), max_val.abs()).clamp(min=1e-12)
|
||||
scale = finfo.max / amax * 0.1
|
||||
x_scl_sat = (x * scale).clamp(min=finfo.min, max=finfo.max)
|
||||
return x_scl_sat.to(dtype), scale.float().reciprocal()
|
||||
|
||||
|
||||
@torch.no_grad()
|
||||
def benchmark_decode(
|
||||
dtype: torch.dtype,
|
||||
quant_dtypes: tuple[
|
||||
Optional[torch.dtype], Optional[torch.dtype], Optional[torch.dtype]
|
||||
],
|
||||
batch_size: int,
|
||||
max_seq_len: int,
|
||||
num_heads: tuple[int, int] = (64, 8),
|
||||
head_size: int = 128,
|
||||
kv_layout: str = "HND",
|
||||
block_size: int = 16,
|
||||
warmup: int = 10,
|
||||
trials: int = 20,
|
||||
):
|
||||
torch.set_default_device("cuda")
|
||||
torch.manual_seed(0)
|
||||
|
||||
q_quant_dtype, kv_quant_dtype, o_quant_dtype = quant_dtypes
|
||||
q_quant_dtype = q_quant_dtype or dtype
|
||||
kv_quant_dtype = kv_quant_dtype or dtype
|
||||
o_quant_dtype = o_quant_dtype or dtype
|
||||
|
||||
num_qo_heads, num_kv_heads = num_heads
|
||||
assert num_qo_heads % num_kv_heads == 0
|
||||
|
||||
sm_scale = float(1.0 / (head_size**0.5))
|
||||
|
||||
# large number to reduce kv_cache reuse
|
||||
NUM_BLOCKS = int(256000 / block_size)
|
||||
|
||||
kv_cache_shape = None
|
||||
if kv_layout == "NHD":
|
||||
kv_cache_shape = (NUM_BLOCKS, 2, block_size, num_kv_heads, head_size)
|
||||
elif kv_layout == "HND":
|
||||
kv_cache_shape = (NUM_BLOCKS, 2, num_kv_heads, block_size, head_size)
|
||||
else:
|
||||
raise ValueError(f"Invalid kv_layout: {kv_layout}")
|
||||
|
||||
# Always using 1.0 scale to reflect the real perf in benchmarking
|
||||
q_scale = 1.0
|
||||
ref_query = torch.randn(batch_size, num_qo_heads, head_size, dtype=dtype)
|
||||
if q_quant_dtype == FP8_DTYPE:
|
||||
query, _ = to_float8(ref_query)
|
||||
else:
|
||||
query = ref_query
|
||||
|
||||
kv_lens = torch.randint(1, max_seq_len, (batch_size,), dtype=torch.int32)
|
||||
kv_lens[-1] = max_seq_len
|
||||
|
||||
seq_lens = kv_lens
|
||||
max_seq_len = torch.max(seq_lens).item()
|
||||
|
||||
# Always using 1.0 scale to reflect the real perf in benchmarking
|
||||
k_scale = v_scale = 1.0
|
||||
ref_kv_cache = torch.randn(kv_cache_shape, dtype=dtype)
|
||||
if kv_quant_dtype == FP8_DTYPE:
|
||||
kv_cache, _ = to_float8(ref_kv_cache)
|
||||
else:
|
||||
kv_cache = ref_kv_cache
|
||||
|
||||
max_num_blocks_per_seq = (max_seq_len + block_size - 1) // block_size
|
||||
block_tables = torch.randint(
|
||||
0, NUM_BLOCKS, (batch_size, max_num_blocks_per_seq), dtype=torch.int32
|
||||
)
|
||||
kv_indptr = [0]
|
||||
kv_indices = []
|
||||
kv_last_page_lens = []
|
||||
for i in range(batch_size):
|
||||
seq_len = seq_lens[i]
|
||||
assert seq_len > 0
|
||||
num_blocks = (seq_len + block_size - 1) // block_size
|
||||
kv_indices.extend(block_tables[i, :num_blocks])
|
||||
kv_indptr.append(kv_indptr[-1] + num_blocks)
|
||||
kv_last_page_len = seq_len % block_size
|
||||
if kv_last_page_len == 0:
|
||||
kv_last_page_len = block_size
|
||||
kv_last_page_lens.append(kv_last_page_len)
|
||||
|
||||
kv_indptr = torch.tensor(kv_indptr, dtype=torch.int32)
|
||||
kv_indices = torch.tensor(kv_indices, dtype=torch.int32)
|
||||
kv_last_page_lens = torch.tensor(kv_last_page_lens, dtype=torch.int32)
|
||||
workspace_buffer = torch.zeros(1024 * 1024 * 1024, dtype=torch.int8)
|
||||
|
||||
wrapper = flashinfer.BatchDecodeWithPagedKVCacheWrapper(
|
||||
workspace_buffer,
|
||||
kv_layout,
|
||||
use_tensor_cores=True,
|
||||
)
|
||||
wrapper.plan(
|
||||
kv_indptr,
|
||||
kv_indices,
|
||||
kv_last_page_lens,
|
||||
num_qo_heads,
|
||||
num_kv_heads,
|
||||
head_size,
|
||||
block_size,
|
||||
"NONE",
|
||||
sm_scale=sm_scale,
|
||||
q_data_type=dtype,
|
||||
kv_data_type=dtype,
|
||||
)
|
||||
|
||||
def time_fn(fn, warmup=10, trials=20):
|
||||
torch.cuda.synchronize()
|
||||
start = torch.cuda.Event(enable_timing=True)
|
||||
end = torch.cuda.Event(enable_timing=True)
|
||||
times = []
|
||||
for i in range(warmup):
|
||||
fn()
|
||||
for i in range(trials):
|
||||
start.record()
|
||||
fn()
|
||||
end.record()
|
||||
torch.cuda.synchronize()
|
||||
times.append(start.elapsed_time(end)) # ms
|
||||
return sum(times) / len(times), torch.std(torch.tensor(times))
|
||||
|
||||
o_scale = 1.0
|
||||
o_sf_scale = None
|
||||
output_baseline = torch.empty(ref_query.shape, dtype=dtype)
|
||||
if o_quant_dtype == FP4_DTYPE:
|
||||
o_sf_scale = 500.0
|
||||
output_trtllm = flashinfer.utils.FP4Tensor(
|
||||
torch.empty(query.shape[:-1] + (query.shape[-1] // 2,), dtype=torch.uint8),
|
||||
torch.empty(
|
||||
(
|
||||
round_up(query.shape[0], 128),
|
||||
round_up(query.shape[1] * query.shape[2] // 16, 4),
|
||||
),
|
||||
dtype=torch.float8_e4m3fn,
|
||||
),
|
||||
)
|
||||
else:
|
||||
output_trtllm = torch.empty(query.shape, dtype=o_quant_dtype)
|
||||
|
||||
def baseline_decode():
|
||||
return wrapper.run(
|
||||
ref_query,
|
||||
ref_kv_cache,
|
||||
k_scale=k_scale,
|
||||
v_scale=v_scale,
|
||||
out=output_baseline,
|
||||
)
|
||||
|
||||
def trtllm_decode():
|
||||
return flashinfer.decode.trtllm_batch_decode_with_kv_cache(
|
||||
query=query,
|
||||
kv_cache=kv_cache,
|
||||
workspace_buffer=workspace_buffer,
|
||||
block_tables=block_tables,
|
||||
seq_lens=seq_lens,
|
||||
max_seq_len=max_seq_len,
|
||||
bmm1_scale=q_scale * k_scale * sm_scale,
|
||||
bmm2_scale=v_scale / o_scale,
|
||||
o_sf_scale=o_sf_scale,
|
||||
out=output_trtllm,
|
||||
)
|
||||
|
||||
baseline_mean, baseline_std = time_fn(baseline_decode)
|
||||
trtllm_mean, trtllm_std = time_fn(trtllm_decode)
|
||||
|
||||
# Calculate percentage speedup (positive means TRT is faster)
|
||||
speedup_percent = (baseline_mean - trtllm_mean) / baseline_mean
|
||||
|
||||
print(
|
||||
f"\t{batch_size}\t{max_seq_len}\t{trtllm_mean:.3f}\t{trtllm_std.item():.3f}"
|
||||
f"\t{baseline_mean:.3f}\t{baseline_std.item():.3f}\t{speedup_percent:.3f}"
|
||||
)
|
||||
|
||||
# Return results for CSV writing
|
||||
return {
|
||||
"batch_size": batch_size,
|
||||
"trtllm_mean": trtllm_mean,
|
||||
"trtllm_std": trtllm_std.item(),
|
||||
"baseline_mean": baseline_mean,
|
||||
"baseline_std": baseline_std.item(),
|
||||
"speedup_percent": speedup_percent,
|
||||
"q_dtype": str(q_quant_dtype),
|
||||
"kv_cache_dtype": str(kv_quant_dtype),
|
||||
"output_dtype": str(o_quant_dtype),
|
||||
"block_size": block_size,
|
||||
"num_kv_heads": num_kv_heads,
|
||||
"head_size": head_size,
|
||||
"max_seq_len": max_seq_len,
|
||||
}
|
||||
|
||||
|
||||
def write_results_to_csv(results, filename=None):
|
||||
"""Write benchmark results to CSV file."""
|
||||
if filename is None:
|
||||
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
||||
filename = f"flashinfer_trtllm_benchmark_{timestamp}.csv"
|
||||
|
||||
fieldnames = [
|
||||
"batch_size",
|
||||
"trtllm_mean",
|
||||
"trtllm_std",
|
||||
"baseline_mean",
|
||||
"baseline_std",
|
||||
"speedup_percent",
|
||||
"q_dtype",
|
||||
"kv_cache_dtype",
|
||||
"output_dtype",
|
||||
"block_size",
|
||||
"num_kv_heads",
|
||||
"head_size",
|
||||
"max_seq_len",
|
||||
]
|
||||
|
||||
file_exists = os.path.exists(filename)
|
||||
|
||||
with open(filename, "a", newline="") as csvfile:
|
||||
writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
|
||||
|
||||
if not file_exists:
|
||||
writer.writeheader()
|
||||
|
||||
for result in results:
|
||||
writer.writerow(result)
|
||||
|
||||
print(f"Results written to {filename}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
batch_sizes = [1, 4, 8, 16, 32, 64, 128, 256]
|
||||
max_seq_lens = [1024, 2048, 4096, 8192, 16384, 32768, 65536, 131072]
|
||||
all_results = []
|
||||
|
||||
dtype = torch.bfloat16
|
||||
quant_dtypes = [
|
||||
# (q_quant_dtype, kv_quant_dtype, o_quant_dtype)
|
||||
(None, None, None),
|
||||
(None, FP8_DTYPE, None),
|
||||
(FP8_DTYPE, FP8_DTYPE, FP8_DTYPE),
|
||||
(FP8_DTYPE, FP8_DTYPE, FP4_DTYPE),
|
||||
]
|
||||
|
||||
for quant_dtype in quant_dtypes:
|
||||
q_quant_dtype, kv_quant_dtype, o_quant_dtype = quant_dtype
|
||||
q_quant_dtype = q_quant_dtype or dtype
|
||||
kv_quant_dtype = kv_quant_dtype or dtype
|
||||
o_quant_dtype = o_quant_dtype or dtype
|
||||
|
||||
print(
|
||||
f"Running benchmark for q_dtype = {q_quant_dtype}, "
|
||||
f"kv_cache_dtype: {kv_quant_dtype}, "
|
||||
f"output_dtype: {o_quant_dtype}"
|
||||
)
|
||||
print(
|
||||
"\tbatch_size\tmax_seq_len\ttrtllm_mean\ttrtllm_std\tbaseline_mean\t"
|
||||
"baseline_std\tspeedup_percent"
|
||||
)
|
||||
for max_seq_len in max_seq_lens:
|
||||
for bs in batch_sizes:
|
||||
result = benchmark_decode(
|
||||
dtype=dtype,
|
||||
quant_dtypes=quant_dtype,
|
||||
batch_size=bs,
|
||||
max_seq_len=max_seq_len,
|
||||
)
|
||||
all_results.append(result)
|
||||
|
||||
# Write all results to CSV
|
||||
write_results_to_csv(all_results)
|
||||
@ -1,307 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import csv
|
||||
import os
|
||||
from datetime import datetime
|
||||
from typing import Optional
|
||||
|
||||
import flashinfer
|
||||
import torch
|
||||
|
||||
from vllm.utils import round_up
|
||||
|
||||
FLOAT32_BYTES = torch.finfo(torch.float).bits // 8
|
||||
FP8_DTYPE = torch.float8_e4m3fn
|
||||
FP4_DTYPE = torch.uint8
|
||||
|
||||
|
||||
def to_float8(x, dtype=torch.float8_e4m3fn):
|
||||
finfo = torch.finfo(dtype)
|
||||
min_val, max_val = x.aminmax()
|
||||
amax = torch.maximum(min_val.abs(), max_val.abs()).clamp(min=1e-12)
|
||||
scale = finfo.max / amax * 0.1
|
||||
x_scl_sat = (x * scale).clamp(min=finfo.min, max=finfo.max)
|
||||
return x_scl_sat.to(dtype), scale.float().reciprocal()
|
||||
|
||||
|
||||
@torch.no_grad()
|
||||
def benchmark_prefill(
|
||||
dtype: torch.dtype,
|
||||
quant_dtypes: tuple[
|
||||
Optional[torch.dtype], Optional[torch.dtype], Optional[torch.dtype]
|
||||
],
|
||||
batch_size: int,
|
||||
max_seq_len: int,
|
||||
num_heads: tuple[int, int] = (64, 8),
|
||||
head_size: int = 128,
|
||||
kv_layout: str = "HND",
|
||||
block_size: int = 16,
|
||||
warmup: int = 10,
|
||||
trials: int = 20,
|
||||
):
|
||||
torch.set_default_device("cuda")
|
||||
torch.manual_seed(0)
|
||||
|
||||
q_quant_dtype, kv_quant_dtype, o_quant_dtype = quant_dtypes
|
||||
q_quant_dtype = q_quant_dtype or dtype
|
||||
kv_quant_dtype = kv_quant_dtype or dtype
|
||||
o_quant_dtype = o_quant_dtype or dtype
|
||||
|
||||
max_q_len = max_kv_len = max_seq_len
|
||||
|
||||
num_qo_heads, num_kv_heads = num_heads
|
||||
assert num_qo_heads % num_kv_heads == 0
|
||||
|
||||
sm_scale = float(1.0 / (head_size**0.5))
|
||||
|
||||
# large number to reduce kv_cache reuse
|
||||
NUM_BLOCKS = int(256000 / block_size)
|
||||
|
||||
kv_cache_shape = None
|
||||
if kv_layout == "NHD":
|
||||
kv_cache_shape = (NUM_BLOCKS, 2, block_size, num_kv_heads, head_size)
|
||||
elif kv_layout == "HND":
|
||||
kv_cache_shape = (NUM_BLOCKS, 2, num_kv_heads, block_size, head_size)
|
||||
else:
|
||||
raise ValueError(f"Invalid kv_layout: {kv_layout}")
|
||||
|
||||
q_lens = torch.randint(1, max_q_len, (batch_size,), dtype=torch.int32)
|
||||
q_lens[-1] = max_q_len
|
||||
q_indptr = torch.cat(
|
||||
[
|
||||
torch.tensor([0], dtype=torch.int32),
|
||||
torch.cumsum(q_lens, dim=0, dtype=torch.int32),
|
||||
]
|
||||
)
|
||||
|
||||
# Always using 1.0 scale to reflect the real perf in benchmarking
|
||||
q_scale = 1.0
|
||||
ref_query = torch.randn(
|
||||
torch.sum(q_lens).item(), num_qo_heads, head_size, dtype=dtype
|
||||
)
|
||||
if q_quant_dtype == FP8_DTYPE:
|
||||
query, _ = to_float8(ref_query)
|
||||
else:
|
||||
query = ref_query
|
||||
|
||||
kv_lens = torch.randint(0, max_kv_len, (batch_size,), dtype=torch.int32)
|
||||
kv_lens[-1] = max_kv_len
|
||||
|
||||
seq_lens = kv_lens + q_lens
|
||||
max_seq_len = torch.max(seq_lens).item()
|
||||
|
||||
# Always using 1.0 scale to reflect the real perf in benchmarking
|
||||
k_scale = v_scale = 1.0
|
||||
ref_kv_cache = torch.randn(kv_cache_shape, dtype=dtype)
|
||||
if kv_quant_dtype == FP8_DTYPE:
|
||||
kv_cache, _ = to_float8(ref_kv_cache)
|
||||
else:
|
||||
kv_cache = ref_kv_cache
|
||||
|
||||
max_num_blocks_per_seq = (max_seq_len + block_size - 1) // block_size
|
||||
block_tables = torch.randint(
|
||||
0, NUM_BLOCKS, (batch_size, max_num_blocks_per_seq), dtype=torch.int32
|
||||
)
|
||||
kv_indptr = [0]
|
||||
kv_indices = []
|
||||
kv_last_page_lens = []
|
||||
for i in range(batch_size):
|
||||
seq_len = seq_lens[i]
|
||||
assert seq_len > 0
|
||||
num_blocks = (seq_len + block_size - 1) // block_size
|
||||
kv_indices.extend(block_tables[i, :num_blocks])
|
||||
kv_indptr.append(kv_indptr[-1] + num_blocks)
|
||||
kv_last_page_len = seq_len % block_size
|
||||
if kv_last_page_len == 0:
|
||||
kv_last_page_len = block_size
|
||||
kv_last_page_lens.append(kv_last_page_len)
|
||||
|
||||
kv_indptr = torch.tensor(kv_indptr, dtype=torch.int32)
|
||||
kv_indices = torch.tensor(kv_indices, dtype=torch.int32)
|
||||
kv_last_page_lens = torch.tensor(kv_last_page_lens, dtype=torch.int32)
|
||||
workspace_buffer = torch.zeros(1024 * 1024 * 1024, dtype=torch.int8)
|
||||
|
||||
wrapper = flashinfer.BatchPrefillWithPagedKVCacheWrapper(
|
||||
workspace_buffer, kv_layout
|
||||
)
|
||||
wrapper.plan(
|
||||
q_indptr,
|
||||
kv_indptr,
|
||||
kv_indices,
|
||||
kv_last_page_lens,
|
||||
num_qo_heads,
|
||||
num_kv_heads,
|
||||
head_size,
|
||||
block_size,
|
||||
causal=True,
|
||||
sm_scale=sm_scale,
|
||||
q_data_type=dtype,
|
||||
kv_data_type=dtype,
|
||||
)
|
||||
|
||||
def time_fn(fn, warmup=10, trials=20):
|
||||
torch.cuda.synchronize()
|
||||
start = torch.cuda.Event(enable_timing=True)
|
||||
end = torch.cuda.Event(enable_timing=True)
|
||||
times = []
|
||||
for i in range(warmup):
|
||||
fn()
|
||||
for i in range(trials):
|
||||
start.record()
|
||||
fn()
|
||||
end.record()
|
||||
torch.cuda.synchronize()
|
||||
times.append(start.elapsed_time(end)) # ms
|
||||
return sum(times) / len(times), torch.std(torch.tensor(times))
|
||||
|
||||
o_scale = 1.0
|
||||
o_sf_scale = None
|
||||
output_baseline = torch.empty(ref_query.shape, dtype=dtype)
|
||||
if o_quant_dtype == FP4_DTYPE:
|
||||
o_sf_scale = 500.0
|
||||
output_trtllm = flashinfer.utils.FP4Tensor(
|
||||
torch.empty(query.shape[:-1] + (query.shape[-1] // 2,), dtype=torch.uint8),
|
||||
torch.empty(
|
||||
(
|
||||
round_up(query.shape[0], 128),
|
||||
round_up(query.shape[1] * query.shape[2] // 16, 4),
|
||||
),
|
||||
dtype=torch.float8_e4m3fn,
|
||||
),
|
||||
)
|
||||
else:
|
||||
output_trtllm = torch.empty(query.shape, dtype=o_quant_dtype)
|
||||
|
||||
def baseline_prefill():
|
||||
return wrapper.run(
|
||||
ref_query,
|
||||
ref_kv_cache,
|
||||
k_scale=k_scale,
|
||||
v_scale=v_scale,
|
||||
out=output_baseline,
|
||||
)
|
||||
|
||||
def trtllm_prefill():
|
||||
return flashinfer.prefill.trtllm_batch_context_with_kv_cache(
|
||||
query=query,
|
||||
kv_cache=kv_cache,
|
||||
workspace_buffer=workspace_buffer,
|
||||
block_tables=block_tables,
|
||||
seq_lens=seq_lens,
|
||||
max_q_len=max_q_len,
|
||||
max_kv_len=max_seq_len,
|
||||
bmm1_scale=q_scale * k_scale * sm_scale,
|
||||
bmm2_scale=v_scale / o_scale,
|
||||
batch_size=batch_size,
|
||||
cum_seq_lens_q=q_indptr,
|
||||
cum_seq_lens_kv=kv_indptr,
|
||||
o_sf_scale=o_sf_scale,
|
||||
out=output_trtllm,
|
||||
)
|
||||
|
||||
baseline_mean, baseline_std = time_fn(baseline_prefill)
|
||||
trtllm_mean, trtllm_std = time_fn(trtllm_prefill)
|
||||
|
||||
# Calculate percentage speedup (positive means TRT is faster)
|
||||
speedup_percent = (baseline_mean - trtllm_mean) / baseline_mean
|
||||
|
||||
print(
|
||||
f"\t{batch_size}\t{max_seq_len}\t{trtllm_mean:8.3f}\t{trtllm_std.item():8.3f}"
|
||||
f"\t{baseline_mean:8.3f}\t{baseline_std.item():8.3f}\t{speedup_percent:8.3f}"
|
||||
)
|
||||
|
||||
# Return results for CSV writing
|
||||
return {
|
||||
"batch_size": batch_size,
|
||||
"trtllm_mean": trtllm_mean,
|
||||
"trtllm_std": trtllm_std.item(),
|
||||
"baseline_mean": baseline_mean,
|
||||
"baseline_std": baseline_std.item(),
|
||||
"speedup_percent": speedup_percent,
|
||||
"q_dtype": str(q_quant_dtype),
|
||||
"kv_cache_dtype": str(kv_quant_dtype),
|
||||
"output_dtype": str(o_quant_dtype),
|
||||
"block_size": block_size,
|
||||
"num_kv_heads": num_kv_heads,
|
||||
"head_size": head_size,
|
||||
"max_seq_len": max_seq_len,
|
||||
}
|
||||
|
||||
|
||||
def write_results_to_csv(results, filename=None):
|
||||
"""Write benchmark results to CSV file."""
|
||||
if filename is None:
|
||||
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
||||
filename = f"flashinfer_trtllm_benchmark_{timestamp}.csv"
|
||||
|
||||
fieldnames = [
|
||||
"batch_size",
|
||||
"trtllm_mean",
|
||||
"trtllm_std",
|
||||
"baseline_mean",
|
||||
"baseline_std",
|
||||
"speedup_percent",
|
||||
"q_dtype",
|
||||
"kv_cache_dtype",
|
||||
"output_dtype",
|
||||
"block_size",
|
||||
"num_kv_heads",
|
||||
"head_size",
|
||||
"max_seq_len",
|
||||
]
|
||||
|
||||
file_exists = os.path.exists(filename)
|
||||
|
||||
with open(filename, "a", newline="") as csvfile:
|
||||
writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
|
||||
|
||||
if not file_exists:
|
||||
writer.writeheader()
|
||||
|
||||
for result in results:
|
||||
writer.writerow(result)
|
||||
|
||||
print(f"Results written to {filename}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
batch_sizes = [1, 4, 8, 16, 32, 64, 128, 256]
|
||||
max_seq_lens = [1024, 2048, 4096, 8192, 16384, 32768, 65536, 131072]
|
||||
all_results = []
|
||||
|
||||
dtype = torch.bfloat16
|
||||
quant_dtypes = [
|
||||
# (q_quant_dtype, kv_quant_dtype, o_quant_dtype)
|
||||
(None, None, None),
|
||||
(FP8_DTYPE, FP8_DTYPE, FP8_DTYPE),
|
||||
(FP8_DTYPE, FP8_DTYPE, FP4_DTYPE),
|
||||
]
|
||||
|
||||
for quant_dtype in quant_dtypes:
|
||||
q_quant_dtype, kv_quant_dtype, o_quant_dtype = quant_dtype
|
||||
q_quant_dtype = q_quant_dtype or dtype
|
||||
kv_quant_dtype = kv_quant_dtype or dtype
|
||||
o_quant_dtype = o_quant_dtype or dtype
|
||||
|
||||
print(
|
||||
f"Running benchmark for q_dtype = {q_quant_dtype}, "
|
||||
f"kv_cache_dtype: {kv_quant_dtype}, "
|
||||
f"output_dtype: {o_quant_dtype}"
|
||||
)
|
||||
print(
|
||||
"\tbatch_size\tmax_seq_len\ttrtllm_mean\ttrtllm_std\tbaseline_mean\t"
|
||||
"baseline_std\tspeedup_percent"
|
||||
)
|
||||
for max_seq_len in max_seq_lens:
|
||||
for bs in batch_sizes:
|
||||
result = benchmark_prefill(
|
||||
dtype=dtype,
|
||||
quant_dtypes=quant_dtype,
|
||||
batch_size=bs,
|
||||
max_seq_len=max_seq_len,
|
||||
)
|
||||
all_results.append(result)
|
||||
|
||||
# Write all results to CSV
|
||||
write_results_to_csv(all_results)
|
||||
@ -11,8 +11,8 @@ from datetime import datetime
|
||||
from typing import Any
|
||||
|
||||
import torch
|
||||
import tqdm
|
||||
import triton
|
||||
from tqdm import tqdm
|
||||
|
||||
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
|
||||
_w8a8_block_fp8_matmul,
|
||||
@ -141,7 +141,6 @@ def get_weight_shapes(tp_size):
|
||||
# cannot TP
|
||||
total = [
|
||||
(512 + 64, 7168),
|
||||
(2112, 7168),
|
||||
((128 + 64) * 128, 7168),
|
||||
(128 * (128 + 128), 512),
|
||||
(7168, 16384),
|
||||
|
||||
@ -8,7 +8,7 @@ Currently this just includes dense GEMMs and only works on Hopper GPUs.
|
||||
|
||||
You need to install vLLM in your usual fashion, then install DeepGEMM from source in its own directory:
|
||||
|
||||
```bash
|
||||
```
|
||||
git clone --recursive https://github.com/deepseek-ai/DeepGEMM
|
||||
cd DeepGEMM
|
||||
python setup.py install
|
||||
@ -17,7 +17,7 @@ uv pip install -e .
|
||||
|
||||
## Usage
|
||||
|
||||
```console
|
||||
```
|
||||
python benchmark_fp8_block_dense_gemm.py
|
||||
INFO 02-26 21:55:13 [__init__.py:207] Automatically detected platform cuda.
|
||||
===== STARTING FP8 GEMM BENCHMARK =====
|
||||
|
||||
@ -4,16 +4,49 @@
|
||||
# ruff: noqa: E501
|
||||
import time
|
||||
|
||||
# Import DeepGEMM functions
|
||||
import deep_gemm
|
||||
import torch
|
||||
from deep_gemm import calc_diff, ceil_div, get_col_major_tma_aligned_tensor
|
||||
|
||||
# Import vLLM functions
|
||||
from vllm import _custom_ops as ops
|
||||
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
|
||||
get_col_major_tma_aligned_tensor,
|
||||
per_token_group_quant_fp8,
|
||||
w8a8_block_fp8_matmul,
|
||||
)
|
||||
from vllm.triton_utils import triton
|
||||
from vllm.utils.deep_gemm import calc_diff, fp8_gemm_nt, per_block_cast_to_fp8
|
||||
|
||||
|
||||
# Copied from
|
||||
# https://github.com/deepseek-ai/DeepGEMM/blob/78cacf70d41d15d688bd493ebc85845f7f2a3d5d/tests/test_core.py#L9
|
||||
def per_token_cast_to_fp8(
|
||||
x: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""Convert tensor to FP8 format with per-token scaling."""
|
||||
assert x.dim() == 2 and x.size(1) % 128 == 0
|
||||
m, n = x.shape
|
||||
x_view = x.view(m, -1, 128)
|
||||
x_amax = x_view.abs().float().amax(dim=2).view(m, -1).clamp(1e-4)
|
||||
return (x_view * (448.0 / x_amax.unsqueeze(2))).to(
|
||||
torch.float8_e4m3fn).view(m, n), (x_amax / 448.0).view(m, -1)
|
||||
|
||||
|
||||
# Copied from
|
||||
# https://github.com/deepseek-ai/DeepGEMM/blob/78cacf70d41d15d688bd493ebc85845f7f2a3d5d/tests/test_core.py#L17
|
||||
def per_block_cast_to_fp8(
|
||||
x: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""Convert tensor to FP8 format with per-block scaling."""
|
||||
assert x.dim() == 2
|
||||
m, n = x.shape
|
||||
x_padded = torch.zeros((ceil_div(m, 128) * 128, ceil_div(n, 128) * 128),
|
||||
dtype=x.dtype,
|
||||
device=x.device)
|
||||
x_padded[:m, :n] = x
|
||||
x_view = x_padded.view(-1, 128, x_padded.size(1) // 128, 128)
|
||||
x_amax = x_view.abs().float().amax(dim=(1, 3), keepdim=True).clamp(1e-4)
|
||||
x_scaled = (x_view * (448.0 / x_amax)).to(torch.float8_e4m3fn)
|
||||
return x_scaled.view_as(x_padded)[:m, :n].contiguous(), (
|
||||
x_amax / 448.0).view(x_view.size(0), x_view.size(2))
|
||||
|
||||
|
||||
def benchmark_shape(m: int,
|
||||
@ -36,14 +69,14 @@ def benchmark_shape(m: int,
|
||||
|
||||
# Pre-quantize B for all implementations
|
||||
# (weights can be pre-quantized offline)
|
||||
B_deepgemm, B_scale_deepgemm = per_block_cast_to_fp8(B, [128, 128], use_ue8m0=True)
|
||||
B_vllm, B_scale_vllm = per_block_cast_to_fp8(B, [128, 128], use_ue8m0=True)
|
||||
B_deepgemm, B_scale_deepgemm = per_block_cast_to_fp8(B)
|
||||
B_vllm, B_scale_vllm = per_block_cast_to_fp8(B)
|
||||
|
||||
# Block size configuration
|
||||
block_size = [128, 128]
|
||||
|
||||
# Pre-quantize A for all implementations
|
||||
A_deepgemm, A_scale_deepgemm = per_token_group_quant_fp8(A, block_size[1])
|
||||
A_deepgemm, A_scale_deepgemm = per_token_cast_to_fp8(A)
|
||||
A_scale_deepgemm = get_col_major_tma_aligned_tensor(A_scale_deepgemm)
|
||||
C_deepgemm = torch.empty((m, n), device='cuda', dtype=torch.bfloat16)
|
||||
A_vllm, A_scale_vllm = per_token_group_quant_fp8(A, block_size[1])
|
||||
@ -52,7 +85,7 @@ def benchmark_shape(m: int,
|
||||
|
||||
# === DeepGEMM Implementation ===
|
||||
def deepgemm_gemm():
|
||||
fp8_gemm_nt((A_deepgemm, A_scale_deepgemm),
|
||||
deep_gemm.gemm_fp8_fp8_bf16_nt((A_deepgemm, A_scale_deepgemm),
|
||||
(B_deepgemm, B_scale_deepgemm),
|
||||
C_deepgemm)
|
||||
return C_deepgemm
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user