mirror of
https://github.com/vllm-project/vllm.git
synced 2025-10-20 23:03:52 +08:00
Compare commits
1 Commits
main
...
fix-precom
| Author | SHA1 | Date | |
|---|---|---|---|
| e17250f0d2 |
@ -5,11 +5,11 @@ import os
|
||||
import sys
|
||||
import zipfile
|
||||
|
||||
# Read the VLLM_MAX_SIZE_MB environment variable, defaulting to 500 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", 500))
|
||||
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"))
|
||||
)
|
||||
|
||||
@ -1,12 +0,0 @@
|
||||
# For hf script, without -t option (tensor parallel size).
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-chartqa-vllm-vlm-baseline.sh -m meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 -l 100 -t 8
|
||||
model_name: "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8"
|
||||
backend: "vllm-vlm"
|
||||
tasks:
|
||||
- name: "chartqa"
|
||||
metrics:
|
||||
- name: "relaxed_accuracy,none"
|
||||
# TODO(zhewenl): model card is 0.90, but the actual score is 0.80.
|
||||
value: 0.80
|
||||
limit: 100
|
||||
num_fewshot: 0
|
||||
@ -1,10 +0,0 @@
|
||||
# For hf script, without -t option (tensor parallel size).
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-mmlupro-vllm-baseline.sh -m meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 -l 250 -t 8 -f 5
|
||||
model_name: "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8"
|
||||
tasks:
|
||||
- name: "mmlu_pro"
|
||||
metrics:
|
||||
- name: "exact_match,custom-extract"
|
||||
value: 0.80
|
||||
limit: 250 # will run on 250 * 14 subjects = 3500 samples
|
||||
num_fewshot: 5
|
||||
@ -1,5 +1,4 @@
|
||||
# For vllm script, with -t option (tensor parallel size)
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-gsm-vllm-baseline.sh -m RedHatAI/Qwen2.5-VL-3B-Instruct-FP8-Dynamic -l 1319 -t 1
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-gsm-vllm-baseline.sh -m RedHatAI/Qwen2.5-VL-3B-Instruct-FP8-Dynamic -b auto -l 1319 -f 5 -t 1
|
||||
model_name: "RedHatAI/Qwen2.5-VL-3B-Instruct-FP8-Dynamic"
|
||||
tasks:
|
||||
- name: "gsm8k"
|
||||
|
||||
@ -1,12 +0,0 @@
|
||||
# For vllm script, with -t option (tensor parallel size).
|
||||
# bash .buildkite/lm-eval-harness/run-lm-eval-chartqa-vllm-vlm-baseline.sh -m Qwen/Qwen2.5-VL-7B-Instruct -l 2500 -t 1
|
||||
|
||||
model_name: "Qwen/Qwen2.5-VL-7B-Instruct"
|
||||
backend: "vllm-vlm"
|
||||
tasks:
|
||||
- name: "chartqa"
|
||||
metrics:
|
||||
- name: "relaxed_accuracy,none"
|
||||
value: 0.855
|
||||
limit: 2500
|
||||
num_fewshot: 0
|
||||
@ -1 +0,0 @@
|
||||
Meta-Llama-4-Maverick-17B-128E-Instruct-FP8.yaml
|
||||
@ -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
|
||||
|
||||
@ -1 +0,0 @@
|
||||
Meta-Llama-4-Maverick-17B-128E-Instruct-FP8-MM.yaml
|
||||
@ -1 +0,0 @@
|
||||
Qwen2.5-VL-7B-Instruct.yaml
|
||||
@ -1,44 +0,0 @@
|
||||
#!/bin/bash
|
||||
# We can use this script to compute baseline accuracy on chartqa for vllm.
|
||||
#
|
||||
# Make sure you have lm-eval-harness installed:
|
||||
# pip install lm-eval==0.4.9
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
echo "Runs lm eval harness on ChartQA using multimodal vllm."
|
||||
echo "This pathway is intended to be used to create baselines for "
|
||||
echo "our correctness tests in vllm's CI."
|
||||
echo
|
||||
echo "usage: ${0} <options>"
|
||||
echo
|
||||
echo " -m - huggingface stub or local directory of the model"
|
||||
echo " -l - limit number of samples to run"
|
||||
echo " -t - tensor parallel size to run at"
|
||||
echo
|
||||
}
|
||||
|
||||
while getopts "m:l:t:" OPT; do
|
||||
case ${OPT} in
|
||||
m )
|
||||
MODEL="$OPTARG"
|
||||
;;
|
||||
l )
|
||||
LIMIT="$OPTARG"
|
||||
;;
|
||||
t )
|
||||
TP_SIZE="$OPTARG"
|
||||
;;
|
||||
\? )
|
||||
usage
|
||||
exit 1
|
||||
;;
|
||||
esac
|
||||
done
|
||||
|
||||
lm_eval --model vllm-vlm \
|
||||
--model_args "pretrained=$MODEL,tensor_parallel_size=$TP_SIZE" \
|
||||
--tasks chartqa \
|
||||
--batch_size auto \
|
||||
--apply_chat_template \
|
||||
--limit $LIMIT
|
||||
2
.buildkite/lm-eval-harness/run-lm-eval-gsm-hf-baseline.sh
Executable file → Normal file
2
.buildkite/lm-eval-harness/run-lm-eval-gsm-hf-baseline.sh
Executable file → Normal file
@ -2,7 +2,7 @@
|
||||
# We can use this script to compute baseline accuracy on GSM for transformers.
|
||||
#
|
||||
# Make sure you have lm-eval-harness installed:
|
||||
# pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d#egg=lm-eval[api]
|
||||
# pip install lm-eval==0.4.4
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
|
||||
@ -3,7 +3,7 @@
|
||||
# We use this for fp8, which HF does not support.
|
||||
#
|
||||
# Make sure you have lm-eval-harness installed:
|
||||
# pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d#egg=lm-eval[api]
|
||||
# pip install lm-eval==0.4.4
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
@ -46,6 +46,6 @@ while getopts "m:b:l:f:t:" OPT; do
|
||||
done
|
||||
|
||||
lm_eval --model vllm \
|
||||
--model_args "pretrained=$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"
|
||||
|
||||
@ -1,50 +0,0 @@
|
||||
#!/bin/bash
|
||||
# We can use this script to compute baseline accuracy on MMLUPRO for vllm.
|
||||
# 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]
|
||||
|
||||
usage() {
|
||||
echo``
|
||||
echo "Runs lm eval harness on MMLU Pro using huggingface transformers."
|
||||
echo "This pathway is intended to be used to create baselines for "
|
||||
echo "our automated nm-test-accuracy workflow"
|
||||
echo
|
||||
echo "usage: ${0} <options>"
|
||||
echo
|
||||
echo " -m - huggingface stub or local directory of the model"
|
||||
echo " -l - limit number of samples to run"
|
||||
echo " -f - number of fewshot samples to use"
|
||||
echo " -t - tensor parallel size to run at"
|
||||
echo
|
||||
}
|
||||
|
||||
while getopts "m:b:l:f:t:" OPT; do
|
||||
case ${OPT} in
|
||||
m )
|
||||
MODEL="$OPTARG"
|
||||
;;
|
||||
b )
|
||||
BATCH_SIZE="$OPTARG"
|
||||
;;
|
||||
l )
|
||||
LIMIT="$OPTARG"
|
||||
;;
|
||||
f )
|
||||
FEWSHOT="$OPTARG"
|
||||
;;
|
||||
t )
|
||||
TP_SIZE="$OPTARG"
|
||||
;;
|
||||
\? )
|
||||
usage
|
||||
exit 1
|
||||
;;
|
||||
esac
|
||||
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" \
|
||||
--tasks mmlu_pro --num_fewshot "$FEWSHOT" --limit "$LIMIT" \
|
||||
--batch_size auto
|
||||
@ -18,28 +18,20 @@ 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)
|
||||
batch_size = eval_config.get("batch_size", "auto")
|
||||
backend = eval_config.get("backend", "vllm")
|
||||
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=backend,
|
||||
model="vllm",
|
||||
model_args=model_args,
|
||||
tasks=[task["name"] for task in eval_config["tasks"]],
|
||||
num_fewshot=eval_config["num_fewshot"],
|
||||
limit=eval_config["limit"],
|
||||
# TODO(yeq): using chat template w/ fewshot_as_multiturn is supposed help
|
||||
# text models. however, this is regressing measured strict-match for
|
||||
# existing text models in CI, so only apply it for mm.
|
||||
apply_chat_template=backend == "vllm-vlm",
|
||||
batch_size=batch_size,
|
||||
batch_size="auto",
|
||||
)
|
||||
return results
|
||||
|
||||
|
||||
@ -7,11 +7,11 @@ 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
|
||||
|
||||
**Benchmarking Coverage**: latency, throughput and fix-qps serving on A100 (the support for FP8 benchmark on H100 is coming!) and Intel® Xeon® Processors, with different models.
|
||||
**Benchmarking Coverage**: latency, throughput and fix-qps serving on A100 (the support for FP8 benchmark on H100 is coming!), with different models.
|
||||
|
||||
**Benchmarking Duration**: about 1hr.
|
||||
|
||||
@ -28,34 +28,16 @@ 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.
|
||||
|
||||
Manually Trigger the benchmark
|
||||
|
||||
```bash
|
||||
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).
|
||||
- `THROUGHPUT_JSON`: JSON file to use for the throughout tests. Default value is empty string (use default file).
|
||||
- `REMOTE_HOST`: IP for the remote vLLM service to benchmark. Default value is empty string.
|
||||
- `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 +60,7 @@ 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 +68,13 @@ 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 +86,7 @@ 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 +104,8 @@ 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.
|
||||
|
||||
@ -136,23 +119,6 @@ If you do not see the table, please wait till the benchmark finish running.
|
||||
The json version of the table (together with the json version of the benchmark) will be also attached to the markdown file.
|
||||
The raw benchmarking results (in the format of json files) are in the `Artifacts` tab of the benchmarking.
|
||||
|
||||
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.
|
||||
|
||||
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.
|
||||
`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" />
|
||||
|
||||
## Nightly test details
|
||||
|
||||
See [nightly-descriptions.md](nightly-descriptions.md) for the detailed description on test workload, models and docker containers of benchmarking other llm engines.
|
||||
@ -160,9 +126,9 @@ 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 +138,6 @@ 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
|
||||
```
|
||||
```console
|
||||
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`.
|
||||
|
||||
@ -8,30 +8,30 @@ This benchmark aims to:
|
||||
|
||||
Latest results: [results link](https://blog.vllm.ai/2024/09/05/perf-update.html), scroll to the end.
|
||||
|
||||
Latest reproduction guide: [github issue link](https://github.com/vllm-project/vllm/issues/8176)
|
||||
Latest reproduction guilde: [github issue link](https://github.com/vllm-project/vllm/issues/8176)
|
||||
|
||||
## 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
|
||||
|
||||
|
||||
@ -1,12 +1,10 @@
|
||||
# Performance benchmarks descriptions
|
||||
|
||||
## Latency tests
|
||||
|
||||
- Input length: 32 tokens.
|
||||
- Output length: 128 tokens.
|
||||
- Batch size: fixed (8).
|
||||
- GPU Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- CPU Models: llama-3.1 8B.
|
||||
- Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- Evaluation metrics: end-to-end latency (mean, median, p99).
|
||||
|
||||
{latency_tests_markdown_table}
|
||||
@ -16,8 +14,7 @@
|
||||
- Input length: randomly sample 200 prompts from ShareGPT dataset (with fixed random seed).
|
||||
- Output length: the corresponding output length of these 200 prompts.
|
||||
- Batch size: dynamically determined by vllm to achieve maximum throughput.
|
||||
- GPU Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- CPU Models: llama-3.1 8B.
|
||||
- Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- Evaluation metrics: throughput.
|
||||
|
||||
{throughput_tests_markdown_table}
|
||||
@ -28,18 +25,12 @@
|
||||
- Output length: the corresponding output length of these 200 prompts.
|
||||
- Batch size: dynamically determined by vllm and the arrival pattern of the requests.
|
||||
- **Average QPS (query per second)**: 1, 4, 16 and inf. QPS = inf means all requests come at once. For other QPS values, the arrival time of each query is determined using a random Poisson process (with fixed random seed).
|
||||
- GPU Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- We also added a speculative decoding test for llama-3 70B on GPU, under QPS 2
|
||||
- CPU Models: llama-3.1 8B.
|
||||
- Models: llama-3.1 8B, llama-3 70B, mixtral 8x7B.
|
||||
- We also added a speculative decoding test for llama-3 70B, under QPS 2
|
||||
- Evaluation metrics: throughput, TTFT (time to the first token, with mean, median and p99), ITL (inter-token latency, with mean, median and p99).
|
||||
- For CPU, we added random dataset tests to benchmark fixed input/output length with 100 prompts.
|
||||
|
||||
{serving_tests_markdown_table}
|
||||
|
||||
## Platform Information
|
||||
|
||||
{platform_markdown_table}
|
||||
|
||||
## json version of the benchmarking tables
|
||||
|
||||
This section contains the data of the markdown tables above in JSON format.
|
||||
|
||||
@ -1,307 +0,0 @@
|
||||
# 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
|
||||
):
|
||||
"""
|
||||
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)
|
||||
|
||||
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)
|
||||
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)
|
||||
|
||||
# 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
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument(
|
||||
"-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",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
|
||||
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
|
||||
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(
|
||||
files,
|
||||
name_column,
|
||||
data_cols_to_compare[i],
|
||||
info_cols,
|
||||
drop_column,
|
||||
debug=debug,
|
||||
)
|
||||
|
||||
# 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"))
|
||||
@ -1,19 +1,15 @@
|
||||
# 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 = {
|
||||
@ -33,39 +29,28 @@ throughput_results = []
|
||||
throughput_results_column_mapping = {
|
||||
"test_name": "Test name",
|
||||
"gpu_type": "GPU",
|
||||
"num_requests": "# of req.",
|
||||
"total_num_tokens": "Total # of tokens",
|
||||
"elapsed_time": "Elapsed time (s)",
|
||||
# "num_requests": "# of req.",
|
||||
# "total_num_tokens": "Total # of tokens",
|
||||
# "elapsed_time": "Elapsed time (s)",
|
||||
"requests_per_second": "Tput (req/s)",
|
||||
"tokens_per_second": "Tput (tok/s)",
|
||||
# "tokens_per_second": "Tput (tok/s)",
|
||||
}
|
||||
|
||||
# serving results and the keys that will be printed into markdown
|
||||
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.",
|
||||
# "completed": "# of req.",
|
||||
"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",
|
||||
# "input_throughput": "Input Tput (tok/s)",
|
||||
# "output_throughput": "Output Tput (tok/s)",
|
||||
"mean_ttft_ms": "Mean TTFT (ms)",
|
||||
"median_ttft_ms": "Median TTFT (ms)",
|
||||
"p99_ttft_ms": "P99 TTFT (ms)",
|
||||
"mean_tpot_ms": "Mean TPOT (ms)",
|
||||
"median_tpot_ms": "Median",
|
||||
"p99_tpot_ms": "P99",
|
||||
# "mean_tpot_ms": "Mean TPOT (ms)",
|
||||
# "median_tpot_ms": "Median",
|
||||
# "p99_tpot_ms": "P99",
|
||||
"mean_itl_ms": "Mean ITL (ms)",
|
||||
"median_itl_ms": "Median ITL (ms)",
|
||||
"p99_itl_ms": "P99 ITL (ms)",
|
||||
@ -90,125 +75,15 @@ def results_to_json(latency, throughput, serving):
|
||||
)
|
||||
|
||||
|
||||
def get_size_with_unit(bytes, suffix="B"):
|
||||
"""
|
||||
Scale bytes to its proper format
|
||||
e.g:
|
||||
1253656 => '1.20MB'
|
||||
1253656678 => '1.17GB'
|
||||
"""
|
||||
factor = 1024
|
||||
for unit in ["", "K", "M", "G", "T", "P"]:
|
||||
if bytes < factor:
|
||||
return f"{bytes:.2f}{unit}{suffix}"
|
||||
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 +91,18 @@ 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 +130,7 @@ 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:
|
||||
@ -312,27 +155,6 @@ if __name__ == "__main__":
|
||||
serving_results = pd.DataFrame.from_dict(serving_results)
|
||||
throughput_results = pd.DataFrame.from_dict(throughput_results)
|
||||
|
||||
svmem = psutil.virtual_memory()
|
||||
platform_data = {
|
||||
"Physical cores": [psutil.cpu_count(logical=False)],
|
||||
"Total cores": [psutil.cpu_count(logical=True)],
|
||||
"Total Memory": [get_size_with_unit(svmem.total)],
|
||||
}
|
||||
|
||||
if util.find_spec("numa") is not None:
|
||||
from numa import info
|
||||
|
||||
platform_data["Total NUMA nodes"] = [info.get_num_configured_nodes()]
|
||||
|
||||
if util.find_spec("cpuinfo") is not None:
|
||||
from cpuinfo import get_cpu_info
|
||||
|
||||
platform_data["CPU Brand"] = [get_cpu_info()["brand_raw"]]
|
||||
|
||||
platform_results = pd.DataFrame.from_dict(
|
||||
platform_data, orient="index", columns=["Platform Info"]
|
||||
)
|
||||
|
||||
raw_results_json = results_to_json(
|
||||
latency_results, throughput_results, serving_results
|
||||
)
|
||||
@ -343,10 +165,7 @@ 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:
|
||||
@ -368,7 +187,7 @@ if __name__ == "__main__":
|
||||
# The GPUs sometimes come in format of "GPUTYPE\nGPUTYPE\n...",
|
||||
# we want to turn it into "8xGPUTYPE"
|
||||
df["GPU"] = df["GPU"].apply(
|
||||
lambda x: f"{len(x.splitlines())}x{x.splitlines()[0]}"
|
||||
lambda x: f"{len(x.split('\n'))}x{x.split('\n')[0]}"
|
||||
)
|
||||
|
||||
# get markdown tables
|
||||
@ -381,14 +200,9 @@ if __name__ == "__main__":
|
||||
throughput_md_table = tabulate(
|
||||
throughput_results, headers="keys", tablefmt="pipe", showindex=False
|
||||
)
|
||||
platform_md_table = tabulate(
|
||||
platform_results, headers="keys", tablefmt="pipe", showindex=True
|
||||
)
|
||||
|
||||
# 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"
|
||||
@ -397,13 +211,12 @@ if __name__ == "__main__":
|
||||
latency_tests_markdown_table=latency_md_table,
|
||||
throughput_tests_markdown_table=throughput_md_table,
|
||||
serving_tests_markdown_table=serving_md_table,
|
||||
platform_markdown_table=platform_md_table,
|
||||
benchmarking_results_in_json_string=processed_results_json,
|
||||
)
|
||||
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")
|
||||
|
||||
@ -181,14 +181,18 @@ launch_vllm_server() {
|
||||
if echo "$common_params" | jq -e 'has("fp8")' >/dev/null; then
|
||||
echo "Key 'fp8' exists in common params. Use neuralmagic fp8 model for convenience."
|
||||
model=$(echo "$common_params" | jq -r '.neuralmagic_quantized_model')
|
||||
server_command="vllm serve $model \
|
||||
server_command="python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
-tp $tp \
|
||||
--model $model \
|
||||
--port $port \
|
||||
$server_args"
|
||||
else
|
||||
echo "Key 'fp8' does not exist in common params."
|
||||
server_command="vllm serve $model \
|
||||
server_command="python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
-tp $tp \
|
||||
--model $model \
|
||||
--port $port \
|
||||
$server_args"
|
||||
fi
|
||||
|
||||
@ -73,7 +73,7 @@ get_current_llm_serving_engine() {
|
||||
echo "Container: vllm"
|
||||
# move to a completely irrelevant directory, to avoid import vllm from current folder
|
||||
export CURRENT_LLM_SERVING_ENGINE=vllm
|
||||
|
||||
|
||||
return
|
||||
fi
|
||||
}
|
||||
@ -95,14 +95,12 @@ json2args() {
|
||||
}
|
||||
|
||||
kill_gpu_processes() {
|
||||
pkill -f '[p]ython'
|
||||
pkill -f '[p]ython3'
|
||||
pkill -f '[t]ritonserver'
|
||||
pkill -f '[p]t_main_thread'
|
||||
pkill -f '[t]ext-generation'
|
||||
pkill -f '[l]mdeploy'
|
||||
# vLLM now names the process with VLLM prefix after https://github.com/vllm-project/vllm/pull/21445
|
||||
pkill -f '[V]LLM'
|
||||
pkill -f python
|
||||
pkill -f python3
|
||||
pkill -f tritonserver
|
||||
pkill -f pt_main_thread
|
||||
pkill -f text-generation
|
||||
pkill -f lmdeploy
|
||||
|
||||
while [ "$(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() {
|
||||
|
||||
@ -31,20 +31,6 @@ check_gpus() {
|
||||
echo "GPU type is $gpu_type"
|
||||
}
|
||||
|
||||
check_cpus() {
|
||||
# check the number of CPUs and NUMA Node and GPU type.
|
||||
declare -g numa_count=$(lscpu | grep "NUMA node(s):" | awk '{print $3}')
|
||||
if [[ $numa_count -gt 0 ]]; then
|
||||
echo "NUMA found."
|
||||
echo $numa_count
|
||||
else
|
||||
echo "Need at least 1 NUMA to run benchmarking."
|
||||
exit 1
|
||||
fi
|
||||
declare -g gpu_type="cpu"
|
||||
echo "GPU type is $gpu_type"
|
||||
}
|
||||
|
||||
check_hf_token() {
|
||||
# check if HF_TOKEN is available and valid
|
||||
if [[ -z "$HF_TOKEN" ]]; then
|
||||
@ -83,22 +69,6 @@ json2args() {
|
||||
echo "$args"
|
||||
}
|
||||
|
||||
json2envs() {
|
||||
# transforms the JSON string to environment variables.
|
||||
# example:
|
||||
# input: { "VLLM_CPU_KVCACHE_SPACE": 5 }
|
||||
# output: VLLM_CPU_KVCACHE_SPACE=5
|
||||
local json_string=$1
|
||||
local args=$(
|
||||
echo "$json_string" | jq -r '
|
||||
to_entries |
|
||||
map((.key ) + "=" + (.value | tostring)) |
|
||||
join(" ")
|
||||
'
|
||||
)
|
||||
echo "$args"
|
||||
}
|
||||
|
||||
wait_for_server() {
|
||||
# wait for vllm server to start
|
||||
# return 1 if vllm server crashes
|
||||
@ -126,8 +96,7 @@ 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 +134,7 @@ upload_to_buildkite() {
|
||||
}
|
||||
|
||||
run_latency_tests() {
|
||||
# run latency tests using `vllm bench latency` command
|
||||
# run latency tests using `benchmark_latency.py`
|
||||
# $1: a json file specifying latency test cases
|
||||
|
||||
local latency_test_file
|
||||
@ -189,26 +158,15 @@ run_latency_tests() {
|
||||
# get arguments
|
||||
latency_params=$(echo "$params" | jq -r '.parameters')
|
||||
latency_args=$(json2args "$latency_params")
|
||||
latency_environment_variables=$(echo "$params" | jq -r '.environment_variables')
|
||||
latency_envs=$(json2envs "$latency_environment_variables")
|
||||
|
||||
# check if there is enough GPU to run the test
|
||||
tp=$(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."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
|
||||
latency_command=" $latency_envs vllm bench latency \
|
||||
latency_command="python3 benchmark_latency.py \
|
||||
--output-json $RESULTS_FOLDER/${test_name}.json \
|
||||
$latency_args"
|
||||
|
||||
@ -234,7 +192,7 @@ run_latency_tests() {
|
||||
}
|
||||
|
||||
run_throughput_tests() {
|
||||
# run throughput tests using `vllm bench throughput`
|
||||
# run throughput tests using `benchmark_throughput.py`
|
||||
# $1: a json file specifying throughput test cases
|
||||
|
||||
local throughput_test_file
|
||||
@ -258,26 +216,15 @@ run_throughput_tests() {
|
||||
# get arguments
|
||||
throughput_params=$(echo "$params" | jq -r '.parameters')
|
||||
throughput_args=$(json2args "$throughput_params")
|
||||
throughput_environment_variables=$(echo "$params" | jq -r '.environment_variables')
|
||||
throughput_envs=$(json2envs "$throughput_environment_variables")
|
||||
|
||||
# check if there is enough GPU to run the test
|
||||
tp=$(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."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
|
||||
throughput_command=" $throughput_envs vllm bench throughput \
|
||||
throughput_command="python3 benchmark_throughput.py \
|
||||
--output-json $RESULTS_FOLDER/${test_name}.json \
|
||||
$throughput_args"
|
||||
|
||||
@ -302,7 +249,7 @@ run_throughput_tests() {
|
||||
}
|
||||
|
||||
run_serving_tests() {
|
||||
# run serving tests using `vllm bench serve` command
|
||||
# run serving tests using `benchmark_serving.py`
|
||||
# $1: a json file specifying serving test cases
|
||||
|
||||
local serving_test_file
|
||||
@ -325,36 +272,18 @@ run_serving_tests() {
|
||||
|
||||
# get client and server arguments
|
||||
server_params=$(echo "$params" | jq -r '.server_parameters')
|
||||
server_envs=$(echo "$params" | jq -r '.server_environment_variables')
|
||||
client_params=$(echo "$params" | jq -r '.client_parameters')
|
||||
server_args=$(json2args "$server_params")
|
||||
server_envs=$(json2envs "$server_envs")
|
||||
client_args=$(json2args "$client_params")
|
||||
qps_list=$(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 $max_concurrency_list"
|
||||
|
||||
# check if there is enough resources to run the test
|
||||
# check if there is enough GPU to run the test
|
||||
tp=$(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."
|
||||
continue
|
||||
fi
|
||||
else
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
if [[ $gpu_count -lt $tp ]]; then
|
||||
echo "Required tensor-parallel-size $tp but only $gpu_count GPU found. Skip testcase $test_name."
|
||||
continue
|
||||
fi
|
||||
|
||||
# check if server model and client model is aligned
|
||||
@ -365,32 +294,23 @@ run_serving_tests() {
|
||||
continue
|
||||
fi
|
||||
|
||||
server_command="$server_envs vllm serve \
|
||||
server_command="python3 \
|
||||
-m vllm.entrypoints.openai.api_server \
|
||||
$server_args"
|
||||
|
||||
# run the server
|
||||
echo "Running test case $test_name"
|
||||
echo "Server command: $server_command"
|
||||
# support remote vllm server
|
||||
client_remote_args=""
|
||||
if [[ -z "${REMOTE_HOST}" ]]; then
|
||||
bash -c "$server_command" &
|
||||
server_pid=$!
|
||||
# wait until the server is alive
|
||||
if wait_for_server; then
|
||||
echo ""
|
||||
echo "vLLM server is up and running."
|
||||
else
|
||||
echo ""
|
||||
echo "vLLM failed to start within the timeout period."
|
||||
fi
|
||||
bash -c "$server_command" &
|
||||
server_pid=$!
|
||||
|
||||
# wait until the server is alive
|
||||
if wait_for_server; then
|
||||
echo ""
|
||||
echo "vllm server is up and running."
|
||||
else
|
||||
server_command="Using Remote Server $REMOTE_HOST $REMOTE_PORT"
|
||||
if [[ ${REMOTE_PORT} ]]; then
|
||||
client_remote_args=" --host=$REMOTE_HOST --port=$REMOTE_PORT "
|
||||
else
|
||||
client_remote_args=" --host=$REMOTE_HOST "
|
||||
fi
|
||||
echo ""
|
||||
echo "vllm failed to start within the timeout period."
|
||||
fi
|
||||
|
||||
# iterate over different QPS
|
||||
@ -402,39 +322,35 @@ 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" \
|
||||
$client_args"
|
||||
|
||||
bash -c "$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
|
||||
@ -444,22 +360,20 @@ run_serving_tests() {
|
||||
}
|
||||
|
||||
main() {
|
||||
local ARCH
|
||||
ARCH=''
|
||||
if [ "$ON_CPU" == "1" ];then
|
||||
check_cpus
|
||||
ARCH='-cpu'
|
||||
else
|
||||
check_gpus
|
||||
fi
|
||||
check_gpus
|
||||
check_hf_token
|
||||
|
||||
# Set to v1 to run v1 benchmark
|
||||
if [[ "${ENGINE_VERSION:-v0}" == "v1" ]]; then
|
||||
export VLLM_USE_V1=1
|
||||
fi
|
||||
|
||||
# dependencies
|
||||
(which wget && which curl) || (apt-get update && apt-get install -y wget curl)
|
||||
(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"
|
||||
@ -472,9 +386,9 @@ main() {
|
||||
QUICK_BENCHMARK_ROOT=../.buildkite/nightly-benchmarks/
|
||||
|
||||
# benchmarking
|
||||
run_serving_tests $QUICK_BENCHMARK_ROOT/tests/"${SERVING_JSON:-serving-tests$ARCH.json}"
|
||||
run_latency_tests $QUICK_BENCHMARK_ROOT/tests/"${LATENCY_JSON:-latency-tests$ARCH.json}"
|
||||
run_throughput_tests $QUICK_BENCHMARK_ROOT/tests/"${THROUGHPUT_JSON:-throughput-tests$ARCH.json}"
|
||||
run_serving_tests $QUICK_BENCHMARK_ROOT/tests/serving-tests.json
|
||||
run_latency_tests $QUICK_BENCHMARK_ROOT/tests/latency-tests.json
|
||||
run_throughput_tests $QUICK_BENCHMARK_ROOT/tests/throughput-tests.json
|
||||
|
||||
# postprocess benchmarking results
|
||||
pip install tabulate pandas
|
||||
|
||||
@ -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"
|
||||
},
|
||||
|
||||
@ -1,30 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "latency_llama8B_tp1",
|
||||
"environment_variables": {
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"load_format": "dummy",
|
||||
"num_iters_warmup": 5,
|
||||
"num_iters": 15
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "latency_llama8B_tp4",
|
||||
"environment_variables": {
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"load_format": "dummy",
|
||||
"num_iters_warmup": 5,
|
||||
"num_iters": 15
|
||||
}
|
||||
}
|
||||
]
|
||||
@ -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"
|
||||
},
|
||||
|
||||
@ -1,610 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "serving_llama8B_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
}
|
||||
]
|
||||
@ -1,820 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "serving_llama8B_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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_bf16_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
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int8_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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"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": "RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "sharegpt",
|
||||
"dataset_path": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "serving_llama8B_int4_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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"quantization": "awq",
|
||||
"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": "hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4",
|
||||
"backend": "vllm",
|
||||
"dataset_name": "random",
|
||||
"random-input-len": 128,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"num_prompts": 1000
|
||||
}
|
||||
}
|
||||
]
|
||||
@ -1,168 +0,0 @@
|
||||
[
|
||||
{
|
||||
"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",
|
||||
"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": [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",
|
||||
"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": [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",
|
||||
"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_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",
|
||||
"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": 1024,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"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",
|
||||
"pipeline_parallel_size": 6,
|
||||
"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": 1024,
|
||||
"random-output-len": 128,
|
||||
"ignore-eos": "",
|
||||
"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": {
|
||||
|
||||
@ -1,32 +0,0 @@
|
||||
[
|
||||
{
|
||||
"test_name": "throughput_llama8B_tp1",
|
||||
"environment_variables": {
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 1,
|
||||
"load_format": "dummy",
|
||||
"dataset": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200,
|
||||
"backend": "vllm"
|
||||
}
|
||||
},
|
||||
{
|
||||
"test_name": "throughput_llama8B_tp4",
|
||||
"environment_variables": {
|
||||
"VLLM_ALLOW_LONG_MAX_MODEL_LEN": 1,
|
||||
"VLLM_CPU_KVCACHE_SPACE": 40
|
||||
},
|
||||
"parameters": {
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"tensor_parallel_size": 4,
|
||||
"load_format": "dummy",
|
||||
"dataset": "./ShareGPT_V3_unfiltered_cleaned_split.json",
|
||||
"num_prompts": 200,
|
||||
"backend": "vllm"
|
||||
}
|
||||
}
|
||||
]
|
||||
46
.buildkite/pyproject.toml
Normal file
46
.buildkite/pyproject.toml
Normal file
@ -0,0 +1,46 @@
|
||||
# This local pyproject file is part of the migration from yapf to ruff format.
|
||||
# It uses the same core rules as the main pyproject.toml file, but with the
|
||||
# following differences:
|
||||
# - ruff line length is overridden to 88
|
||||
# - deprecated typing ignores (UP006, UP035) have been removed
|
||||
|
||||
[tool.ruff]
|
||||
line-length = 88
|
||||
|
||||
[tool.ruff.lint.per-file-ignores]
|
||||
"vllm/third_party/**" = ["ALL"]
|
||||
"vllm/version.py" = ["F401"]
|
||||
"vllm/_version.py" = ["ALL"]
|
||||
|
||||
[tool.ruff.lint]
|
||||
select = [
|
||||
# pycodestyle
|
||||
"E",
|
||||
# Pyflakes
|
||||
"F",
|
||||
# pyupgrade
|
||||
"UP",
|
||||
# flake8-bugbear
|
||||
"B",
|
||||
# flake8-simplify
|
||||
"SIM",
|
||||
# isort
|
||||
"I",
|
||||
# flake8-logging-format
|
||||
"G",
|
||||
]
|
||||
ignore = [
|
||||
# star imports
|
||||
"F405", "F403",
|
||||
# lambda expression assignment
|
||||
"E731",
|
||||
# Loop control variable not used within loop body
|
||||
"B007",
|
||||
# f-string format
|
||||
"UP032",
|
||||
# Can remove once 3.10+ is the minimum Python version
|
||||
"UP007",
|
||||
]
|
||||
|
||||
[tool.ruff.format]
|
||||
docstring-code-format = true
|
||||
@ -1,36 +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"
|
||||
depends_on: ~
|
||||
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 VLLM_MAIN_CUDA_VERSION=12.9 --build-arg torch_cuda_arch_list='8.7 8.9 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"
|
||||
|
||||
# aarch64 build.
|
||||
- label: "Build arm64 CPU wheel"
|
||||
depends_on: ~
|
||||
id: build-wheel-arm64-cpu
|
||||
agents:
|
||||
queue: arm64_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 --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile.cpu ."
|
||||
- "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 wheel - CUDA 12.8"
|
||||
depends_on: ~
|
||||
id: build-wheel-cuda-12-8
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
@ -43,7 +12,6 @@ steps:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- label: "Build wheel - CUDA 12.6"
|
||||
depends_on: ~
|
||||
id: build-wheel-cuda-12-6
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
@ -55,61 +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 --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 FLASHINFER_AOT_COMPILE=true --build-arg torch_cuda_arch_list='8.7 8.9 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-11-8
|
||||
id: annotate-release-workflow
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
@ -150,52 +101,22 @@ steps:
|
||||
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 GIT_REPO_CHECK=1 --build-arg VLLM_CPU_AVX512BF16=true --build-arg VLLM_CPU_AVX512VNNI=true --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:latest --progress plain --target vllm-openai -f docker/Dockerfile.cpu ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:latest"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:latest --progress plain --target vllm-openai -f docker/Dockerfile.cpu ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:$(buildkite-agent meta-data get release-version)"
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- block: "Build arm64 CPU release image"
|
||||
key: block-arm64-cpu-release-image-build
|
||||
- block: "Build Neuron release image"
|
||||
key: block-neuron-release-image-build
|
||||
depends_on: ~
|
||||
|
||||
- label: "Build and publish arm64 CPU release image"
|
||||
depends_on: block-arm64-cpu-release-image-build
|
||||
- label: "Build and publish Neuron release image"
|
||||
depends_on: block-neuron-release-image-build
|
||||
agents:
|
||||
queue: arm64_cpu_queue_postmerge
|
||||
queue: neuron-postmerge
|
||||
commands:
|
||||
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --tag public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:latest --progress plain --target vllm-openai -f docker/Dockerfile.cpu ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:latest"
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:$(buildkite-agent meta-data get release-version)"
|
||||
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --tag public.ecr.aws/q9t5s3a7/vllm-neuron-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-neuron-release-repo:latest --progress plain -f docker/Dockerfile.neuron ."
|
||||
- "docker push public.ecr.aws/q9t5s3a7/vllm-neuron-release-repo:$(buildkite-agent meta-data get release-version)"
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
|
||||
- label: "Build and publish nightly multi-arch image to DockerHub"
|
||||
depends_on:
|
||||
- create-multi-arch-manifest
|
||||
if: build.env("NIGHTLY") == "1"
|
||||
agents:
|
||||
queue: cpu_queue_postmerge
|
||||
commands:
|
||||
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
|
||||
- "docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-x86_64"
|
||||
- "docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-aarch64"
|
||||
- "docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-x86_64 vllm/vllm-openai:nightly-x86_64"
|
||||
- "docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-aarch64 vllm/vllm-openai:nightly-aarch64"
|
||||
- "docker push vllm/vllm-openai:nightly-x86_64"
|
||||
- "docker push vllm/vllm-openai:nightly-aarch64"
|
||||
- "docker manifest create vllm/vllm-openai:nightly vllm/vllm-openai:nightly-x86_64 vllm/vllm-openai:nightly-aarch64 --amend"
|
||||
- "docker manifest create vllm/vllm-openai:nightly-$BUILDKITE_COMMIT vllm/vllm-openai:nightly-x86_64 vllm/vllm-openai:nightly-aarch64 --amend"
|
||||
- "docker manifest push vllm/vllm-openai:nightly"
|
||||
- "docker manifest push vllm/vllm-openai:nightly-$BUILDKITE_COMMIT"
|
||||
# Clean up old nightly builds (keep only last 14)
|
||||
- "bash .buildkite/scripts/cleanup-nightly-builds.sh"
|
||||
plugins:
|
||||
- docker-login#v3.0.0:
|
||||
username: vllmbot
|
||||
password-env: DOCKERHUB_TOKEN
|
||||
env:
|
||||
DOCKER_BUILDKIT: "1"
|
||||
DOCKERHUB_USERNAME: "vllmbot"
|
||||
|
||||
@ -14,33 +14,18 @@ buildkite-agent annotate --style 'info' --context 'release-workflow' << EOF
|
||||
To download the wheel:
|
||||
\`\`\`
|
||||
aws s3 cp s3://vllm-wheels/${RELEASE_VERSION}/vllm-${RELEASE_VERSION}-cp38-abi3-manylinux1_x86_64.whl .
|
||||
aws s3 cp s3://vllm-wheels/${RELEASE_VERSION}/vllm-${RELEASE_VERSION}-cp38-abi3-manylinux2014_aarch64.whl .
|
||||
|
||||
aws s3 cp s3://vllm-wheels/${RELEASE_VERSION}+cu126/vllm-${RELEASE_VERSION}+cu126-cp38-abi3-manylinux1_x86_64.whl .
|
||||
aws s3 cp s3://vllm-wheels/${RELEASE_VERSION}+cu129/vllm-${RELEASE_VERSION}+cu129-cp38-abi3-manylinux1_x86_64.whl .
|
||||
aws s3 cp s3://vllm-wheels/${RELEASE_VERSION}+cu118/vllm-${RELEASE_VERSION}+cu118-cp38-abi3-manylinux1_x86_64.whl .
|
||||
\`\`\`
|
||||
|
||||
To download and upload the image:
|
||||
|
||||
\`\`\`
|
||||
docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-x86_64
|
||||
docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-aarch64
|
||||
|
||||
docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-x86_64 vllm/vllm-openai:x86_64
|
||||
docker tag vllm/vllm-openai:x86_64 vllm/vllm-openai:latest-x86_64
|
||||
docker tag vllm/vllm-openai:x86_64 vllm/vllm-openai:v${RELEASE_VERSION}-x86_64
|
||||
docker push vllm/vllm-openai:latest-x86_64
|
||||
docker push vllm/vllm-openai:v${RELEASE_VERSION}-x86_64
|
||||
|
||||
docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-aarch64 vllm/vllm-openai:aarch64
|
||||
docker tag vllm/vllm-openai:aarch64 vllm/vllm-openai:latest-aarch64
|
||||
docker tag vllm/vllm-openai:aarch64 vllm/vllm-openai:v${RELEASE_VERSION}-aarch64
|
||||
docker push vllm/vllm-openai:latest-aarch64
|
||||
docker push vllm/vllm-openai:v${RELEASE_VERSION}-aarch64
|
||||
|
||||
docker manifest create vllm/vllm-openai:latest vllm/vllm-openai:latest-x86_64 vllm/vllm-openai:latest-aarch64 --amend
|
||||
docker manifest create vllm/vllm-openai:v${RELEASE_VERSION} vllm/vllm-openai:v${RELEASE_VERSION}-x86_64 vllm/vllm-openai:v${RELEASE_VERSION}-aarch64 --amend
|
||||
docker manifest push vllm/vllm-openai:latest
|
||||
docker manifest push vllm/vllm-openai:v${RELEASE_VERSION}
|
||||
docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}
|
||||
docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT} vllm/vllm-openai
|
||||
docker tag vllm/vllm-openai vllm/vllm-openai:latest
|
||||
docker tag vllm/vllm-openai vllm/vllm-openai:v${RELEASE_VERSION}
|
||||
docker push vllm/vllm-openai:latest
|
||||
docker push vllm/vllm-openai:v${RELEASE_VERSION}
|
||||
\`\`\`
|
||||
EOF
|
||||
@ -1,120 +0,0 @@
|
||||
#!/bin/bash
|
||||
|
||||
set -ex
|
||||
|
||||
# Clean up old nightly builds from DockerHub, keeping only the last 14 builds
|
||||
# This script uses DockerHub API to list and delete old tags with "nightly-" prefix
|
||||
|
||||
# DockerHub API endpoint for vllm/vllm-openai repository
|
||||
REPO_API_URL="https://hub.docker.com/v2/repositories/vllm/vllm-openai/tags"
|
||||
|
||||
# Get DockerHub credentials from environment
|
||||
if [ -z "$DOCKERHUB_TOKEN" ]; then
|
||||
echo "Error: DOCKERHUB_TOKEN environment variable is not set"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
if [ -z "$DOCKERHUB_USERNAME" ]; then
|
||||
echo "Error: DOCKERHUB_USERNAME environment variable is not set"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# Get DockerHub bearer token
|
||||
echo "Getting DockerHub bearer token..."
|
||||
set +x
|
||||
BEARER_TOKEN=$(curl -s -X POST \
|
||||
-H "Content-Type: application/json" \
|
||||
-d "{\"username\": \"$DOCKERHUB_USERNAME\", \"password\": \"$DOCKERHUB_TOKEN\"}" \
|
||||
"https://hub.docker.com/v2/users/login" | jq -r '.token')
|
||||
set -x
|
||||
|
||||
if [ -z "$BEARER_TOKEN" ] || [ "$BEARER_TOKEN" = "null" ]; then
|
||||
echo "Error: Failed to get DockerHub bearer token"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# Function to get all tags from DockerHub
|
||||
get_all_tags() {
|
||||
local page=1
|
||||
local all_tags=""
|
||||
|
||||
while true; do
|
||||
set +x
|
||||
local response=$(curl -s -H "Authorization: Bearer $BEARER_TOKEN" \
|
||||
"$REPO_API_URL?page=$page&page_size=100")
|
||||
set -x
|
||||
|
||||
# Get both last_updated timestamp and tag name, separated by |
|
||||
local tags=$(echo "$response" | jq -r '.results[] | select(.name | startswith("nightly-")) | "\(.last_updated)|\(.name)"')
|
||||
|
||||
if [ -z "$tags" ]; then
|
||||
break
|
||||
fi
|
||||
|
||||
all_tags="$all_tags$tags"$'\n'
|
||||
page=$((page + 1))
|
||||
done
|
||||
|
||||
# Sort by timestamp (newest first) and extract just the tag names
|
||||
echo "$all_tags" | sort -r | cut -d'|' -f2
|
||||
}
|
||||
|
||||
delete_tag() {
|
||||
local tag_name="$1"
|
||||
echo "Deleting tag: $tag_name"
|
||||
|
||||
local delete_url="https://hub.docker.com/v2/repositories/vllm/vllm-openai/tags/$tag_name"
|
||||
set +x
|
||||
local response=$(curl -s -X DELETE -H "Authorization: Bearer $BEARER_TOKEN" "$delete_url")
|
||||
set -x
|
||||
|
||||
if echo "$response" | jq -e '.detail' > /dev/null 2>&1; then
|
||||
echo "Warning: Failed to delete tag $tag_name: $(echo "$response" | jq -r '.detail')"
|
||||
else
|
||||
echo "Successfully deleted tag: $tag_name"
|
||||
fi
|
||||
}
|
||||
|
||||
# Get all nightly- prefixed tags, sorted by last_updated timestamp (newest first)
|
||||
echo "Fetching all tags from DockerHub..."
|
||||
all_tags=$(get_all_tags)
|
||||
|
||||
if [ -z "$all_tags" ]; then
|
||||
echo "No tags found to clean up"
|
||||
exit 0
|
||||
fi
|
||||
|
||||
# Count total tags
|
||||
total_tags=$(echo "$all_tags" | wc -l)
|
||||
echo "Found $total_tags tags"
|
||||
|
||||
# Keep only the last 14 builds (including the current one)
|
||||
tags_to_keep=14
|
||||
tags_to_delete=$((total_tags - tags_to_keep))
|
||||
|
||||
if [ $tags_to_delete -le 0 ]; then
|
||||
echo "No tags need to be deleted (only $total_tags tags found, keeping $tags_to_keep)"
|
||||
exit 0
|
||||
fi
|
||||
|
||||
echo "Will delete $tags_to_delete old tags, keeping the newest $tags_to_keep"
|
||||
|
||||
# Get tags to delete (skip the first $tags_to_keep tags)
|
||||
tags_to_delete_list=$(echo "$all_tags" | tail -n +$((tags_to_keep + 1)))
|
||||
|
||||
if [ -z "$tags_to_delete_list" ]; then
|
||||
echo "No tags to delete"
|
||||
exit 0
|
||||
fi
|
||||
|
||||
# Delete old tags
|
||||
echo "Deleting old tags..."
|
||||
while IFS= read -r tag; do
|
||||
if [ -n "$tag" ]; then
|
||||
delete_tag "$tag"
|
||||
# Add a small delay to avoid rate limiting
|
||||
sleep 1
|
||||
fi
|
||||
done <<< "$tags_to_delete_list"
|
||||
|
||||
echo "Cleanup completed successfully"
|
||||
@ -86,6 +86,10 @@ if [[ $commands == *"pytest -v -s models/test_registry.py"* ]]; then
|
||||
commands=${commands//"pytest -v -s models/test_registry.py"/"pytest -v -s models/test_registry.py -k 'not BambaForCausalLM and not GritLM and not Mamba2ForCausalLM and not Zamba2ForCausalLM'"}
|
||||
fi
|
||||
|
||||
if [[ $commands == *"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2'"* ]]; then
|
||||
commands=${commands//"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2'"/"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2 and not BambaForCausalLM and not Gemma2ForCausalLM and not Grok1ModelForCausalLM and not Zamba2ForCausalLM and not Gemma2Model and not GritLM'"}
|
||||
fi
|
||||
|
||||
if [[ $commands == *"pytest -v -s compile/test_basic_correctness.py"* ]]; then
|
||||
commands=${commands//"pytest -v -s compile/test_basic_correctness.py"/"VLLM_USE_TRITON_FLASH_ATTN=0 pytest -v -s compile/test_basic_correctness.py"}
|
||||
fi
|
||||
@ -103,8 +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 \
|
||||
@ -117,6 +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 \
|
||||
@ -160,9 +167,16 @@ 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
|
||||
|
||||
#Obsolete currently
|
||||
##ignore certain Entrypoints/llm tests
|
||||
#if [[ $commands == *" && pytest -v -s entrypoints/llm/test_guided_generate.py"* ]]; then
|
||||
# commands=${commands//" && pytest -v -s entrypoints/llm/test_guided_generate.py"/" "}
|
||||
#fi
|
||||
|
||||
# --ignore=entrypoints/openai/test_encoder_decoder.py \
|
||||
# --ignore=entrypoints/openai/test_embedding.py \
|
||||
# --ignore=entrypoints/openai/test_oot_registration.py
|
||||
|
||||
@ -25,28 +25,25 @@ function cpu_tests() {
|
||||
|
||||
# offline inference
|
||||
podman exec -it "$container_id" bash -c "
|
||||
set -xve
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m" >> $HOME/test_basic.log
|
||||
set -e
|
||||
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m"
|
||||
|
||||
# Run basic model test
|
||||
podman exec -it "$container_id" bash -c "
|
||||
set -evx
|
||||
set -e
|
||||
pip install pytest pytest-asyncio einops peft Pillow soundfile transformers_stream_generator matplotlib
|
||||
pip install sentence-transformers datamodel_code_generator
|
||||
|
||||
# Note: disable Bart until supports V1
|
||||
# pytest -v -s tests/models/language/generation/test_bart.py -m cpu_model
|
||||
pytest -v -s tests/models/language/generation/test_bart.py -m cpu_model
|
||||
pytest -v -s tests/models/language/generation/test_common.py::test_models[False-5-32-openai-community/gpt2]
|
||||
pytest -v -s tests/models/language/generation/test_common.py::test_models[False-5-32-facebook/opt-125m]
|
||||
pytest -v -s tests/models/language/generation/test_common.py::test_models[False-5-32-google/gemma-1.1-2b-it]
|
||||
pytest -v -s tests/models/language/pooling/test_classification.py::test_models[float-jason9693/Qwen2.5-1.5B-apeach]
|
||||
# TODO: Below test case tests/models/language/pooling/test_embedding.py::test_models[True-ssmits/Qwen2-7B-Instruct-embed-base] fails on ppc64le. Disabling it for time being.
|
||||
# pytest -v -s tests/models/language/pooling/test_embedding.py -m cpu_model" >> $HOME/test_rest.log
|
||||
pytest -v -s tests/models/language/pooling/test_embedding.py -m cpu_model"
|
||||
}
|
||||
|
||||
# All of CPU tests are expected to be finished less than 40 mins.
|
||||
|
||||
export container_id
|
||||
export -f cpu_tests
|
||||
timeout 120m bash -c cpu_tests
|
||||
timeout 40m bash -c cpu_tests
|
||||
|
||||
|
||||
@ -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
|
||||
@ -46,74 +45,57 @@ 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
|
||||
|
||||
pytest -x -v -s tests/models/language/generation -m cpu_model
|
||||
VLLM_CPU_SGL_KERNEL=1 pytest -x -v -s tests/models/language/generation -m cpu_model
|
||||
|
||||
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
|
||||
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"
|
||||
|
||||
# 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"
|
||||
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 1h 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; }
|
||||
|
||||
@ -2,32 +2,10 @@
|
||||
|
||||
# This script build the CPU docker image and run the offline inference inside the container.
|
||||
# It serves a sanity check for compilation and basic model usage.
|
||||
set -exuo pipefail
|
||||
set -ex
|
||||
|
||||
# Try building the docker image
|
||||
cat <<EOF | docker build -t hpu-plugin-v1-test-env -f - .
|
||||
FROM gaudi-base-image:latest
|
||||
|
||||
COPY ./ /workspace/vllm
|
||||
|
||||
WORKDIR /workspace/vllm
|
||||
|
||||
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
|
||||
|
||||
# install development dependencies (for testing)
|
||||
RUN python3 -m pip install -e tests/vllm_test_utils
|
||||
|
||||
WORKDIR /workspace/
|
||||
|
||||
RUN git clone https://github.com/vllm-project/vllm-gaudi.git
|
||||
|
||||
RUN ln -s /workspace/vllm/tests && ln -s /workspace/vllm/examples && ln -s /workspace/vllm/benchmarks
|
||||
|
||||
EOF
|
||||
docker build -t hpu-test-env -f docker/Dockerfile.hpu .
|
||||
|
||||
# Setup cleanup
|
||||
# certain versions of HPU software stack have a bug that can
|
||||
@ -36,21 +14,13 @@ EOF
|
||||
# functions, while other platforms only need one remove_docker_container
|
||||
# function.
|
||||
EXITCODE=1
|
||||
remove_docker_containers() { docker rm -f hpu-plugin-v1-test || true; }
|
||||
trap 'remove_docker_containers; exit $EXITCODE;' EXIT
|
||||
remove_docker_containers() { docker rm -f hpu-test || true; docker rm -f hpu-test-tp2 || true; }
|
||||
remove_docker_containers_and_exit() { remove_docker_containers; exit $EXITCODE; }
|
||||
trap remove_docker_containers_and_exit EXIT
|
||||
remove_docker_containers
|
||||
|
||||
echo "Running HPU plugin v1 test"
|
||||
docker run --rm --runtime=habana --name=hpu-plugin-v1-test --network=host \
|
||||
-e HABANA_VISIBLE_DEVICES=all \
|
||||
hpu-plugin-v1-test-env \
|
||||
/bin/bash "/workspace/vllm-gaudi/tests/upstream_tests/ci_tests.sh"
|
||||
# Run the image and launch offline inference
|
||||
docker run --runtime=habana --name=hpu-test --network=host -e HABANA_VISIBLE_DEVICES=all -e VLLM_SKIP_WARMUP=true --entrypoint="" hpu-test-env python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m
|
||||
docker run --runtime=habana --name=hpu-test-tp2 --network=host -e HABANA_VISIBLE_DEVICES=all -e VLLM_SKIP_WARMUP=true --entrypoint="" hpu-test-env python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --tensor-parallel-size 2
|
||||
|
||||
EXITCODE=$?
|
||||
if [ $EXITCODE -eq 0 ]; then
|
||||
echo "Test with basic model passed"
|
||||
else
|
||||
echo "Test with basic model FAILED with exit code: $EXITCODE" >&2
|
||||
fi
|
||||
|
||||
# The trap will handle the container removal and final exit.
|
||||
63
.buildkite/scripts/hardware_ci/run-neuron-test.sh
Normal file
63
.buildkite/scripts/hardware_ci/run-neuron-test.sh
Normal file
@ -0,0 +1,63 @@
|
||||
#!/bin/bash
|
||||
|
||||
# This script build the Neuron docker image and run the API server inside the container.
|
||||
# It serves a sanity check for compilation and basic model usage.
|
||||
set -e
|
||||
set -v
|
||||
|
||||
image_name="neuron/vllm-ci"
|
||||
container_name="neuron_$(tr -dc A-Za-z0-9 < /dev/urandom | head -c 10; echo)"
|
||||
|
||||
HF_CACHE="$(realpath ~)/huggingface"
|
||||
mkdir -p "${HF_CACHE}"
|
||||
HF_MOUNT="/root/.cache/huggingface"
|
||||
HF_TOKEN=$(aws secretsmanager get-secret-value --secret-id "ci/vllm-neuron/hf-token" --region us-west-2 --query 'SecretString' --output text | jq -r .VLLM_NEURON_CI_HF_TOKEN)
|
||||
|
||||
NEURON_COMPILE_CACHE_URL="$(realpath ~)/neuron_compile_cache"
|
||||
mkdir -p "${NEURON_COMPILE_CACHE_URL}"
|
||||
NEURON_COMPILE_CACHE_MOUNT="/root/.cache/neuron_compile_cache"
|
||||
|
||||
# Try building the docker image
|
||||
aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws
|
||||
|
||||
# prune old image and containers to save disk space, and only once a day
|
||||
# by using a timestamp file in tmp.
|
||||
if [ -f /tmp/neuron-docker-build-timestamp ]; then
|
||||
last_build=$(cat /tmp/neuron-docker-build-timestamp)
|
||||
current_time=$(date +%s)
|
||||
if [ $((current_time - last_build)) -gt 86400 ]; 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 -f
|
||||
echo "$current_time" > /tmp/neuron-docker-build-timestamp
|
||||
fi
|
||||
else
|
||||
date "+%s" > /tmp/neuron-docker-build-timestamp
|
||||
fi
|
||||
|
||||
docker build -t "${image_name}" -f docker/Dockerfile.neuron .
|
||||
|
||||
# Setup cleanup
|
||||
remove_docker_container() {
|
||||
docker image rm -f "${image_name}" || true;
|
||||
}
|
||||
trap remove_docker_container EXIT
|
||||
|
||||
# Run the image
|
||||
docker run --rm -it --device=/dev/neuron0 --network bridge \
|
||||
-v "${HF_CACHE}:${HF_MOUNT}" \
|
||||
-e "HF_HOME=${HF_MOUNT}" \
|
||||
-e "HF_TOKEN=${HF_TOKEN}" \
|
||||
-v "${NEURON_COMPILE_CACHE_URL}:${NEURON_COMPILE_CACHE_MOUNT}" \
|
||||
-e "NEURON_COMPILE_CACHE_URL=${NEURON_COMPILE_CACHE_MOUNT}" \
|
||||
--name "${container_name}" \
|
||||
${image_name} \
|
||||
/bin/bash -c "
|
||||
python3 /workspace/vllm/examples/offline_inference/neuron.py;
|
||||
python3 -m pytest /workspace/vllm/tests/neuron/1_core/ -v --capture=tee-sys;
|
||||
for f in /workspace/vllm/tests/neuron/2_core/*.py; do
|
||||
echo 'Running test file: '$f;
|
||||
python3 -m pytest \$f -v --capture=tee-sys;
|
||||
done
|
||||
"
|
||||
@ -1,191 +0,0 @@
|
||||
#!/bin/bash
|
||||
|
||||
# This script build the Ascend NPU docker image and run the offline inference inside the container.
|
||||
# It serves a sanity check for compilation and basic model usage.
|
||||
set -ex
|
||||
|
||||
# Base ubuntu image with basic ascend development libraries and python installed
|
||||
VLLM_ASCEND_REPO="https://github.com/vllm-project/vllm-ascend.git"
|
||||
CONFIG_FILE_REMOTE_PATH="tests/e2e/vllm_interface/vllm_test.cfg"
|
||||
TEST_RUN_CONFIG_FILE="vllm_test.cfg"
|
||||
VLLM_ASCEND_TMP_DIR=
|
||||
# Get the test run configuration file from the vllm-ascend repository
|
||||
fetch_vllm_test_cfg() {
|
||||
VLLM_ASCEND_TMP_DIR=$(mktemp -d)
|
||||
# Ensure that the temporary directory is cleaned up when an exception occurs during configuration file retrieval
|
||||
cleanup() {
|
||||
rm -rf "${VLLM_ASCEND_TMP_DIR}"
|
||||
}
|
||||
trap cleanup EXIT
|
||||
|
||||
GIT_TRACE=1 git clone -v --depth 1 "${VLLM_ASCEND_REPO}" "${VLLM_ASCEND_TMP_DIR}"
|
||||
if [ ! -f "${VLLM_ASCEND_TMP_DIR}/${CONFIG_FILE_REMOTE_PATH}" ]; then
|
||||
echo "Error: file '${CONFIG_FILE_REMOTE_PATH}' does not exist in the warehouse" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# If the file already exists locally, just overwrite it
|
||||
cp "${VLLM_ASCEND_TMP_DIR}/${CONFIG_FILE_REMOTE_PATH}" "${TEST_RUN_CONFIG_FILE}"
|
||||
echo "Copied ${CONFIG_FILE_REMOTE_PATH} to ${TEST_RUN_CONFIG_FILE}"
|
||||
|
||||
# Since the trap will be overwritten later, and when it is executed here, the task of cleaning up resources
|
||||
# when the trap is abnormal has been completed, so the temporary resources are manually deleted here.
|
||||
rm -rf "${VLLM_ASCEND_TMP_DIR}"
|
||||
trap - EXIT
|
||||
}
|
||||
|
||||
# Downloads test run configuration file from a remote URL.
|
||||
# Loads the configuration into the current script environment.
|
||||
get_config() {
|
||||
if [ ! -f "${TEST_RUN_CONFIG_FILE}" ]; then
|
||||
echo "Error: file '${TEST_RUN_CONFIG_FILE}' does not exist in the warehouse" >&2
|
||||
exit 1
|
||||
fi
|
||||
source "${TEST_RUN_CONFIG_FILE}"
|
||||
echo "Base docker image name that get from configuration: ${BASE_IMAGE_NAME}"
|
||||
return 0
|
||||
}
|
||||
|
||||
# get test running configuration.
|
||||
fetch_vllm_test_cfg
|
||||
get_config
|
||||
# Check if the function call was successful. If not, exit the script.
|
||||
if [ $? -ne 0 ]; then
|
||||
exit 1
|
||||
fi
|
||||
|
||||
image_name="npu/vllm-ci:${BUILDKITE_COMMIT}_${EPOCHSECONDS}"
|
||||
container_name="npu_${BUILDKITE_COMMIT}_$(tr -dc A-Za-z0-9 < /dev/urandom | head -c 10; echo)"
|
||||
|
||||
# BUILDKITE_AGENT_NAME format is {hostname}-{agent_idx}-{npu_card_num}cards
|
||||
agent_idx=$(echo "${BUILDKITE_AGENT_NAME}" | awk -F'-' '{print $(NF-1)}')
|
||||
echo "agent_idx: ${agent_idx}"
|
||||
builder_name="cachebuilder${agent_idx}"
|
||||
builder_cache_dir="/mnt/docker-cache${agent_idx}"
|
||||
mkdir -p ${builder_cache_dir}
|
||||
|
||||
# Try building the docker image
|
||||
cat <<EOF | DOCKER_BUILDKIT=1 docker build \
|
||||
--add-host cache-service-vllm.nginx-pypi-cache.svc.cluster.local:${PYPI_CACHE_HOST} \
|
||||
--builder ${builder_name} --cache-from type=local,src=${builder_cache_dir} \
|
||||
--cache-to type=local,dest=${builder_cache_dir},mode=max \
|
||||
--progress=plain --load -t ${image_name} -f - .
|
||||
FROM ${BASE_IMAGE_NAME}
|
||||
|
||||
# Define environments
|
||||
ENV DEBIAN_FRONTEND=noninteractive
|
||||
|
||||
RUN pip config set global.index-url http://cache-service-vllm.nginx-pypi-cache.svc.cluster.local:${PYPI_CACHE_PORT}/pypi/simple && \
|
||||
pip config set global.trusted-host cache-service-vllm.nginx-pypi-cache.svc.cluster.local && \
|
||||
apt-get update -y && \
|
||||
apt-get install -y python3-pip git vim wget net-tools gcc g++ cmake libnuma-dev && \
|
||||
rm -rf /var/cache/apt/* && \
|
||||
rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Install for pytest to make the docker build cache layer always valid
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
pip install pytest>=6.0 modelscope
|
||||
|
||||
WORKDIR /workspace/vllm
|
||||
|
||||
# Install vLLM dependencies in advance. Effect: As long as common.txt remains unchanged, the docker cache layer will be valid.
|
||||
COPY requirements/common.txt /workspace/vllm/requirements/common.txt
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
pip install -r requirements/common.txt
|
||||
|
||||
COPY . .
|
||||
|
||||
# Install vLLM
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
VLLM_TARGET_DEVICE="empty" python3 -m pip install -v -e /workspace/vllm/ --extra-index https://download.pytorch.org/whl/cpu/ && \
|
||||
python3 -m pip uninstall -y triton
|
||||
|
||||
# Install vllm-ascend
|
||||
WORKDIR /workspace
|
||||
ARG VLLM_ASCEND_REPO=https://github.com/vllm-project/vllm-ascend.git
|
||||
ARG VLLM_ASCEND_TAG=main
|
||||
RUN git config --global url."https://gh-proxy.test.osinfra.cn/https://github.com/".insteadOf "https://github.com/" && \
|
||||
git clone --depth 1 \$VLLM_ASCEND_REPO --branch \$VLLM_ASCEND_TAG /workspace/vllm-ascend
|
||||
|
||||
# Install vllm dependencies in advance. Effect: As long as common.txt remains unchanged, the docker cache layer will be valid.
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
pip install -r /workspace/vllm-ascend/requirements.txt
|
||||
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
|
||||
source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
|
||||
source /usr/local/Ascend/nnal/atb/set_env.sh && \
|
||||
export LD_LIBRARY_PATH=\$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
|
||||
python3 -m pip install -v -e /workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/
|
||||
|
||||
ENV VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
ENV VLLM_USE_MODELSCOPE=True
|
||||
|
||||
WORKDIR /workspace/vllm-ascend
|
||||
|
||||
CMD ["/bin/bash"]
|
||||
|
||||
EOF
|
||||
|
||||
# Setup cleanup
|
||||
remove_docker_container() {
|
||||
docker rm -f "${container_name}" || true;
|
||||
docker image rm -f "${image_name}" || true;
|
||||
docker system prune -f || true;
|
||||
}
|
||||
trap remove_docker_container EXIT
|
||||
|
||||
# Generate corresponding --device args based on BUILDKITE_AGENT_NAME
|
||||
# Ascend NPU BUILDKITE_AGENT_NAME format is {hostname}-{agent_idx}-{npu_card_num}cards, and agent_idx starts from 1.
|
||||
# e.g. atlas-a2-001-1-2cards means this is the 1-th agent on atlas-a2-001 host, and it has 2 NPU cards.
|
||||
# returns --device /dev/davinci0 --device /dev/davinci1
|
||||
parse_and_gen_devices() {
|
||||
local input="$1"
|
||||
local index cards_num
|
||||
if [[ "$input" =~ ([0-9]+)-([0-9]+)cards$ ]]; then
|
||||
index="${BASH_REMATCH[1]}"
|
||||
cards_num="${BASH_REMATCH[2]}"
|
||||
else
|
||||
echo "parse error" >&2
|
||||
return 1
|
||||
fi
|
||||
|
||||
local devices=""
|
||||
local i=0
|
||||
while (( i < cards_num )); do
|
||||
local dev_idx=$(((index - 1)*cards_num + i ))
|
||||
devices="$devices --device /dev/davinci${dev_idx}"
|
||||
((i++))
|
||||
done
|
||||
|
||||
# trim leading space
|
||||
devices="${devices#"${devices%%[![:space:]]*}"}"
|
||||
# Output devices: assigned to the caller variable
|
||||
printf '%s' "$devices"
|
||||
}
|
||||
|
||||
devices=$(parse_and_gen_devices "${BUILDKITE_AGENT_NAME}") || exit 1
|
||||
|
||||
# Run the image and execute the Out-Of-Tree (OOT) platform interface test case on Ascend NPU hardware.
|
||||
# This test checks whether the OOT platform interface is functioning properly in conjunction with
|
||||
# the hardware plugin vllm-ascend.
|
||||
model_cache_dir=/mnt/modelscope${agent_idx}
|
||||
mkdir -p ${model_cache_dir}
|
||||
docker run \
|
||||
${devices} \
|
||||
--device /dev/davinci_manager \
|
||||
--device /dev/devmm_svm \
|
||||
--device /dev/hisi_hdc \
|
||||
-v /usr/local/dcmi:/usr/local/dcmi \
|
||||
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
|
||||
-v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
|
||||
-v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
|
||||
-v /etc/ascend_install.info:/etc/ascend_install.info \
|
||||
-v ${model_cache_dir}:/root/.cache/modelscope \
|
||||
--entrypoint="" \
|
||||
--name "${container_name}" \
|
||||
"${image_name}" \
|
||||
bash -c '
|
||||
set -e
|
||||
pytest -v -s tests/e2e/vllm_interface/
|
||||
'
|
||||
@ -1,166 +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 tblib==3.1.0
|
||||
echo "--- Python dependencies installed ---"
|
||||
|
||||
export VLLM_XLA_CHECK_RECOMPILATION=1
|
||||
export VLLM_XLA_CACHE_PATH=
|
||||
|
||||
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
|
||||
@ -61,15 +62,15 @@ 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 tblib==3.1.0
|
||||
&& 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
|
||||
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
|
||||
@ -148,6 +149,16 @@ 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"
|
||||
|
||||
# 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;
|
||||
}
|
||||
@ -23,26 +23,9 @@ 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
|
||||
pip install tblib==3.1.0
|
||||
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 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_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
|
||||
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
|
||||
'
|
||||
|
||||
@ -11,20 +11,20 @@ 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
|
||||
vllm serve meta-llama/Llama-2-7b-chat-hf &
|
||||
python3 -m vllm.entrypoints.openai.api_server --model meta-llama/Llama-2-7b-chat-hf &
|
||||
server_pid=$!
|
||||
wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
|
||||
# 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 \
|
||||
|
||||
@ -1,59 +0,0 @@
|
||||
#!/bin/bash
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
# Setup script for Prime-RL integration tests
|
||||
# This script prepares the environment for running Prime-RL tests with nightly vLLM
|
||||
|
||||
set -euo pipefail
|
||||
|
||||
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
||||
REPO_ROOT="$(cd "${SCRIPT_DIR}/../.." && pwd)"
|
||||
PRIME_RL_REPO="https://github.com/PrimeIntellect-ai/prime-rl.git"
|
||||
PRIME_RL_DIR="${REPO_ROOT}/prime-rl"
|
||||
|
||||
echo "Setting up Prime-RL integration test environment..."
|
||||
|
||||
# Clean up any existing Prime-RL directory
|
||||
if [ -d "${PRIME_RL_DIR}" ]; then
|
||||
echo "Removing existing Prime-RL directory..."
|
||||
rm -rf "${PRIME_RL_DIR}"
|
||||
fi
|
||||
|
||||
# Install UV if not available
|
||||
if ! command -v uv &> /dev/null; then
|
||||
echo "Installing UV package manager..."
|
||||
curl -LsSf https://astral.sh/uv/install.sh | sh
|
||||
source $HOME/.local/bin/env
|
||||
fi
|
||||
|
||||
# Clone Prime-RL repository at specific branch for reproducible tests
|
||||
PRIME_RL_BRANCH="integ-vllm-main"
|
||||
echo "Cloning Prime-RL repository at branch: ${PRIME_RL_BRANCH}..."
|
||||
git clone --branch "${PRIME_RL_BRANCH}" --single-branch "${PRIME_RL_REPO}" "${PRIME_RL_DIR}"
|
||||
cd "${PRIME_RL_DIR}"
|
||||
|
||||
echo "Setting up UV project environment..."
|
||||
export UV_PROJECT_ENVIRONMENT=/usr/local
|
||||
ln -s /usr/bin/python3 /usr/local/bin/python
|
||||
|
||||
# Remove vllm pin from pyproject.toml
|
||||
echo "Removing vllm pin from pyproject.toml..."
|
||||
sed -i '/vllm==/d' pyproject.toml
|
||||
|
||||
# Sync Prime-RL dependencies
|
||||
echo "Installing Prime-RL dependencies..."
|
||||
uv sync --inexact && uv sync --inexact --all-extras
|
||||
|
||||
# Verify installation
|
||||
echo "Verifying installations..."
|
||||
uv run python -c "import vllm; print(f'vLLM version: {vllm.__version__}')"
|
||||
uv run python -c "import prime_rl; print('Prime-RL imported successfully')"
|
||||
|
||||
echo "Prime-RL integration test environment setup complete!"
|
||||
|
||||
echo "Running Prime-RL integration tests..."
|
||||
export WANDB_MODE=offline # this makes this test not require a WANDB_API_KEY
|
||||
uv run pytest -vs tests/integration/test_rl.py -m gpu
|
||||
|
||||
echo "Prime-RL integration tests completed!"
|
||||
@ -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."
|
||||
|
||||
@ -1,11 +1,11 @@
|
||||
# Environment config
|
||||
TEST_NAME=llama8b
|
||||
CONTAINER_NAME=tpu-test
|
||||
CONTAINER_NAME=vllm-tpu
|
||||
|
||||
# vllm config
|
||||
MODEL=meta-llama/Llama-3.1-8B-Instruct
|
||||
MAX_NUM_SEQS=256
|
||||
MAX_NUM_BATCHED_TOKENS=1024
|
||||
MAX_NUM_SEQS=512
|
||||
MAX_NUM_BATCHED_TOKENS=512
|
||||
TENSOR_PARALLEL_SIZE=1
|
||||
MAX_MODEL_LEN=2048
|
||||
DOWNLOAD_DIR=/mnt/disks/persist
|
||||
|
||||
@ -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"
|
||||
@ -56,7 +68,7 @@ docker run \
|
||||
|
||||
echo "run script..."
|
||||
echo
|
||||
docker exec "$CONTAINER_NAME" /bin/bash -c ".buildkite/scripts/tpu/run_bm.sh"
|
||||
docker exec "$CONTAINER_NAME" /bin/bash -c ".buildkite/scripts/hardware_ci/run_bm.sh"
|
||||
|
||||
echo "copy result back..."
|
||||
VLLM_LOG="$LOG_ROOT/$TEST_NAME"_vllm_log.txt
|
||||
|
||||
@ -1,14 +0,0 @@
|
||||
# Environment config
|
||||
TEST_NAME=llama8bw8a8
|
||||
CONTAINER_NAME=tpu-test
|
||||
|
||||
# vllm config
|
||||
MODEL=RedHatAI/Meta-Llama-3.1-8B-Instruct-quantized.w8a8
|
||||
MAX_NUM_SEQS=128
|
||||
MAX_NUM_BATCHED_TOKENS=1024
|
||||
TENSOR_PARALLEL_SIZE=1
|
||||
MAX_MODEL_LEN=2048
|
||||
DOWNLOAD_DIR=/mnt/disks/persist
|
||||
EXPECTED_THROUGHPUT=8.7
|
||||
INPUT_LEN=1800
|
||||
OUTPUT_LEN=128
|
||||
@ -42,8 +42,9 @@ echo "lanching vllm..."
|
||||
echo "logging to $VLLM_LOG"
|
||||
echo
|
||||
|
||||
vllm serve $MODEL \
|
||||
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
|
||||
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
47
.coveragerc
47
.coveragerc
@ -1,47 +0,0 @@
|
||||
[run]
|
||||
# Track the installed vllm package (this is what actually gets imported during tests)
|
||||
# Use wildcard pattern to match the installed location
|
||||
source =
|
||||
vllm
|
||||
*/dist-packages/vllm
|
||||
*/site-packages/vllm
|
||||
omit =
|
||||
*/tests/*
|
||||
*/test_*
|
||||
*/__pycache__/*
|
||||
*/build/*
|
||||
*/dist/*
|
||||
*/vllm.egg-info/*
|
||||
*/third_party/*
|
||||
*/examples/*
|
||||
*/benchmarks/*
|
||||
*/docs/*
|
||||
|
||||
[paths]
|
||||
# Map all possible vllm locations to a canonical "vllm" path
|
||||
# This ensures coverage.combine properly merges data from different test runs
|
||||
source =
|
||||
vllm
|
||||
/vllm-workspace/src/vllm
|
||||
/vllm-workspace/vllm
|
||||
*/site-packages/vllm
|
||||
*/dist-packages/vllm
|
||||
|
||||
[report]
|
||||
exclude_lines =
|
||||
pragma: no cover
|
||||
def __repr__
|
||||
if self.debug:
|
||||
if settings.DEBUG
|
||||
raise AssertionError
|
||||
raise NotImplementedError
|
||||
if 0:
|
||||
if __name__ == .__main__.:
|
||||
class .*\bProtocol\):
|
||||
@(abc\.)?abstractmethod
|
||||
|
||||
[html]
|
||||
directory = htmlcov
|
||||
|
||||
[xml]
|
||||
output = coverage.xml
|
||||
@ -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
|
||||
@ -1,4 +0,0 @@
|
||||
# Migrate from `yapf` & `isort` to `ruff`
|
||||
d6953beb91da4e9c99be4c0a1304a2d24189535c
|
||||
# Convert `Optional[x]` to `x | None` and `Union[x, y]` to `x | y`
|
||||
8fcaaf6a165e661f63fc51be906bc05b0767332f
|
||||
24
.github/.bc-linter.yml
vendored
24
.github/.bc-linter.yml
vendored
@ -1,24 +0,0 @@
|
||||
# doc: https://github.com/pytorch/test-infra/blob/main/tools/stronghold/docs/bc_linter_config.md
|
||||
version: 1
|
||||
paths:
|
||||
# We temporarily disable globally, and will only enable with `annotations.include`
|
||||
# include:
|
||||
# - "vllm/v1/attetion/*.py"
|
||||
# - "vllm/v1/core/*.py"
|
||||
exclude:
|
||||
- "**/*.py"
|
||||
|
||||
scan:
|
||||
functions: true # check free functions and methods
|
||||
classes: true # check classes/dataclasses
|
||||
public_only: true # ignore names starting with "_" at any level
|
||||
|
||||
annotations:
|
||||
include: # decorators that force‑include a symbol
|
||||
- name: "bc_linter_include" # matched by simple name or dotted suffix
|
||||
propagate_to_members: false # for classes, include methods/inner classes
|
||||
exclude: # decorators that force‑exclude a symbol
|
||||
- name: "bc_linter_skip" # matched by simple name or dotted suffix
|
||||
propagate_to_members: true # for classes, exclude methods/inner classes
|
||||
|
||||
excluded_violations: [] # e.g. ["ParameterRenamed", "FieldTypeChanged"]
|
||||
132
.github/CODEOWNERS
vendored
132
.github/CODEOWNERS
vendored
@ -2,127 +2,49 @@
|
||||
# for more info about CODEOWNERS file
|
||||
|
||||
# This lists cover the "core" components of vLLM that require careful review
|
||||
/vllm/attention @LucasWilkinson
|
||||
/vllm/attention/backends/abstract.py @WoosukKwon @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/executor/executor_base.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill @22quinn
|
||||
/vllm/model_executor/layers/fused_moe @mgoin
|
||||
/vllm/model_executor/layers/quantization @mgoin @robertgshaw2-redhat @tlrmchlsmth @yewentao256
|
||||
/vllm/model_executor/layers/mamba @tdoublep
|
||||
/vllm/model_executor/model_loader @22quinn
|
||||
/vllm/multimodal @DarkLight1337 @ywang96 @NickLucche
|
||||
/vllm/core @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/engine/llm_engine.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/vllm/executor/executor_base.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
|
||||
/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
|
||||
/vllm/model_executor/guided_decoding @mgoin @russellb @aarnphm
|
||||
/vllm/multimodal @DarkLight1337 @ywang96
|
||||
/vllm/vllm_flash_attn @LucasWilkinson
|
||||
/vllm/lora @jeejeelee
|
||||
/vllm/reasoning @aarnphm @chaunceyjiang
|
||||
/vllm/entrypoints @aarnphm @chaunceyjiang
|
||||
/vllm/compilation @zou3519 @youkaichao @ProExpertProg
|
||||
/vllm/distributed/kv_transfer @NickLucche @ApostaC
|
||||
CMakeLists.txt @tlrmchlsmth @LucasWilkinson
|
||||
|
||||
# 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/cache.py @simon-mo @WoosukKwon @youkaichao @robertgshaw2-redhat @mgoin @tlrmchlsmth @houseroad @hmellor @yewentao256 @ProExpertProg @heheda12345
|
||||
/vllm/reasoning @aarnphm
|
||||
/vllm/entrypoints @aarnphm
|
||||
CMakeLists.txt @tlrmchlsmth
|
||||
|
||||
# vLLM V1
|
||||
/vllm/v1/attention @LucasWilkinson
|
||||
/vllm/v1/attention/backends/flashinfer.py @mgoin
|
||||
/vllm/v1/attention/backends/triton_attn.py @tdoublep
|
||||
/vllm/v1/core @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat @heheda12345 @ApostaC
|
||||
/vllm/v1/sample @22quinn @houseroad @njhill
|
||||
/vllm/v1/spec_decode @benchislett @luccafong
|
||||
/vllm/v1/structured_output @mgoin @russellb @aarnphm @benchislett
|
||||
/vllm/v1/kv_cache_interface.py @heheda12345
|
||||
/vllm/v1/offloading @ApostaC
|
||||
/vllm/v1 @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat
|
||||
/vllm/v1/structured_output @mgoin @russellb @aarnphm
|
||||
|
||||
# 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 @NickLucche
|
||||
/tests/evals @mgoin
|
||||
/tests/kernels @mgoin @tlrmchlsmth @WoosukKwon @yewentao256
|
||||
/tests/entrypoints @DarkLight1337 @robertgshaw2-redhat @simon-mo @aarnphm
|
||||
/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/multimodal @DarkLight1337 @ywang96 @NickLucche
|
||||
/tests/quantization @mgoin @robertgshaw2-redhat @yewentao256
|
||||
/tests/multi_step @alexm-redhat @comaniac
|
||||
/tests/multimodal @DarkLight1337 @ywang96
|
||||
/tests/prefix_caching @comaniac @KuntaiDu
|
||||
/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/v1/core @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat @heheda12345 @ApostaC
|
||||
/tests/weight_loading @mgoin @youkaichao @yewentao256
|
||||
/tests/weight_loading @mgoin @youkaichao
|
||||
/tests/lora @jeejeelee
|
||||
/tests/models/language/generation/test_hybrid.py @tdoublep
|
||||
/tests/v1/kv_connector/nixl_integration @NickLucche
|
||||
/tests/v1/kv_connector @ApostaC
|
||||
/tests/v1/offloading @ApostaC
|
||||
|
||||
# Transformers backend
|
||||
/vllm/model_executor/models/transformers @hmellor
|
||||
/tests/models/test_transformers.py @hmellor
|
||||
|
||||
# Docs
|
||||
/docs/mkdocs @hmellor
|
||||
/docs/**/*.yml @hmellor
|
||||
/requirements/docs.txt @hmellor
|
||||
.readthedocs.yaml @hmellor
|
||||
/docs @hmellor
|
||||
mkdocs.yaml @hmellor
|
||||
|
||||
# Linting
|
||||
.markdownlint.yaml @hmellor
|
||||
.pre-commit-config.yaml @hmellor
|
||||
/tools/pre_commit @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
|
||||
|
||||
# MTP-specific files
|
||||
/vllm/model_executor/models/deepseek_mtp.py @luccafong
|
||||
|
||||
# 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
|
||||
|
||||
# TPU
|
||||
/vllm/v1/worker/tpu* @NickLucche
|
||||
/vllm/platforms/tpu.py @NickLucche
|
||||
/vllm/v1/sample/tpu @NickLucche
|
||||
/vllm/tests/v1/tpu @NickLucche
|
||||
|
||||
# KVConnector installation files
|
||||
/requirements/kv_connectors.txt @NickLucche
|
||||
|
||||
# Pooling models
|
||||
/examples/*/pooling/ @noooop
|
||||
/tests/models/*/pooling* @noooop
|
||||
/tests/entrypoints/pooling @noooop
|
||||
/vllm/config/pooler.py @noooop
|
||||
/vllm/pooling_params.py @noooop
|
||||
/vllm/model_executor/layers/pooler.py @noooop
|
||||
|
||||
4
.github/ISSUE_TEMPLATE/750-RFC.yml
vendored
4
.github/ISSUE_TEMPLATE/750-RFC.yml
vendored
@ -43,6 +43,10 @@ body:
|
||||
Any other things you would like to mention.
|
||||
validations:
|
||||
required: false
|
||||
- type: markdown
|
||||
attributes:
|
||||
value: >
|
||||
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)
|
||||
|
||||
145
.github/mergify.yml
vendored
145
.github/mergify.yml
vendored
@ -2,7 +2,6 @@ pull_request_rules:
|
||||
- name: label-documentation
|
||||
description: Automatically apply documentation label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^[^/]+\.md$
|
||||
- files~=^docs/
|
||||
@ -11,13 +10,10 @@ pull_request_rules:
|
||||
label:
|
||||
add:
|
||||
- documentation
|
||||
comment:
|
||||
message: "Documentation preview: https://vllm--{{number}}.org.readthedocs.build/en/{{number}}/"
|
||||
|
||||
- name: label-ci-build
|
||||
description: Automatically apply ci/build label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^\.github/
|
||||
- files~=\.buildkite/
|
||||
@ -31,27 +27,9 @@ pull_request_rules:
|
||||
add:
|
||||
- ci/build
|
||||
|
||||
- name: label-deepseek
|
||||
description: Automatically apply deepseek label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^examples/.*deepseek.*\.py
|
||||
- files~=^tests/.*deepseek.*\.py
|
||||
- files~=^vllm/entrypoints/openai/tool_parsers/.*deepseek.*\.py
|
||||
- files~=^vllm/model_executor/models/.*deepseek.*\.py
|
||||
- files~=^vllm/reasoning/.*deepseek.*\.py
|
||||
- files~=^vllm/transformers_utils/.*deepseek.*\.py
|
||||
- title~=(?i)DeepSeek
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- deepseek
|
||||
|
||||
- name: label-frontend
|
||||
description: Automatically apply frontend label
|
||||
conditions:
|
||||
- label != stale
|
||||
- files~=^vllm/entrypoints/
|
||||
actions:
|
||||
label:
|
||||
@ -61,14 +39,12 @@ pull_request_rules:
|
||||
- name: label-llama
|
||||
description: Automatically apply llama label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^examples/.*llama.*\.py
|
||||
- files~=^tests/.*llama.*\.py
|
||||
- files~=^vllm/entrypoints/openai/tool_parsers/llama.*\.py
|
||||
- files~=^vllm/model_executor/models/.*llama.*\.py
|
||||
- files~=^vllm/transformers_utils/configs/.*llama.*\.py
|
||||
- title~=(?i)llama
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
@ -77,84 +53,36 @@ pull_request_rules:
|
||||
- name: label-multi-modality
|
||||
description: Automatically apply multi-modality label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^vllm/multimodal/
|
||||
- files~=^tests/multimodal/
|
||||
- files~=^tests/models/multimodal/
|
||||
- files~=^tests/models/*/audio_language/
|
||||
- files~=^tests/models/*/vision_language/
|
||||
- files=tests/models/test_vision.py
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- multi-modality
|
||||
|
||||
- name: label-new-model
|
||||
description: Automatically apply new-model label
|
||||
conditions:
|
||||
- label != stale
|
||||
- and:
|
||||
- files~=^vllm/model_executor/models/
|
||||
- files=vllm/model_executor/models/registry.py
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- new-model
|
||||
|
||||
- name: label-performance
|
||||
description: Automatically apply performance label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^benchmarks/
|
||||
- files~=^vllm/benchmarks/
|
||||
- files~=^tests/benchmarks/
|
||||
- files~=^\.buildkite/nightly-benchmarks/
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- performance
|
||||
|
||||
- name: label-qwen
|
||||
description: Automatically apply qwen label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^examples/.*qwen.*\.py
|
||||
- files~=^tests/.*qwen.*\.py
|
||||
- files~=^vllm/model_executor/models/.*qwen.*\.py
|
||||
- files~=^vllm/reasoning/.*qwen.*\.py
|
||||
- title~=(?i)Qwen
|
||||
- body~=(?i)Qwen
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- qwen
|
||||
|
||||
- name: label-gpt-oss
|
||||
description: Automatically apply gpt-oss label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^examples/.*gpt[-_]?oss.*\.py
|
||||
- files~=^tests/.*gpt[-_]?oss.*\.py
|
||||
- files~=^tests/entrypoints/openai/test_response_api_with_harmony.py
|
||||
- files~=^tests/entrypoints/test_context.py
|
||||
- files~=^vllm/model_executor/models/.*gpt[-_]?oss.*\.py
|
||||
- files~=^vllm/model_executor/layers/.*gpt[-_]?oss.*\.py
|
||||
- files~=^vllm/entrypoints/harmony_utils.py
|
||||
- files~=^vllm/entrypoints/tool_server.py
|
||||
- files~=^vllm/entrypoints/tool.py
|
||||
- files~=^vllm/entrypoints/context.py
|
||||
- title~=(?i)gpt[-_]?oss
|
||||
- title~=(?i)harmony
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- gpt-oss
|
||||
|
||||
- name: label-rocm
|
||||
description: Automatically apply rocm label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^csrc/rocm/
|
||||
- files~=^docker/Dockerfile.rocm
|
||||
@ -175,7 +103,6 @@ pull_request_rules:
|
||||
- name: label-structured-output
|
||||
description: Automatically apply structured-output label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^benchmarks/structured_schemas/
|
||||
- files=benchmarks/benchmark_serving_structured_output.py
|
||||
@ -184,8 +111,11 @@ 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_struct_output_generate.py
|
||||
- files=tests/v1/entrypoints/llm/test_guided_generate.py
|
||||
- files~=^vllm/v1/structured_output/
|
||||
actions:
|
||||
label:
|
||||
@ -195,14 +125,10 @@ pull_request_rules:
|
||||
- name: label-speculative-decoding
|
||||
description: Automatically apply speculative-decoding label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^vllm/v1/spec_decode/
|
||||
- files~=^tests/v1/spec_decode/
|
||||
- files~=^examples/.*(spec_decode|mlpspeculator|eagle|speculation).*\.py
|
||||
- files~=^vllm/model_executor/models/.*eagle.*\.py
|
||||
- files=vllm/model_executor/models/mlp_speculator.py
|
||||
- files~=^vllm/transformers_utils/configs/(eagle|medusa|mlp_speculator)\.py
|
||||
- files~=^vllm/spec_decode/
|
||||
- files=vllm/model_executor/layers/spec_decode_base_sampler.py
|
||||
- files~=^tests/spec_decode/
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
@ -211,7 +137,6 @@ pull_request_rules:
|
||||
- name: label-v1
|
||||
description: Automatically apply v1 label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^vllm/v1/
|
||||
- files~=^tests/v1/
|
||||
@ -224,7 +149,6 @@ pull_request_rules:
|
||||
description: Automatically apply tpu label
|
||||
# Keep this list in sync with `label-tpu-remove` conditions
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=tpu.py
|
||||
- files~=_tpu
|
||||
@ -240,7 +164,6 @@ pull_request_rules:
|
||||
description: Automatically remove tpu label
|
||||
# Keep this list in sync with `label-tpu` conditions
|
||||
conditions:
|
||||
- label != stale
|
||||
- and:
|
||||
- -files~=tpu.py
|
||||
- -files~=_tpu
|
||||
@ -255,9 +178,9 @@ pull_request_rules:
|
||||
- name: label-tool-calling
|
||||
description: Automatically add tool-calling label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^tests/tool_use/
|
||||
- files~=^tests/mistral_tool_use/
|
||||
- files~=^tests/entrypoints/openai/tool_parsers/
|
||||
- files=tests/entrypoints/openai/test_chat_with_tool_reasoning.py
|
||||
- files~=^vllm/entrypoints/openai/tool_parsers/
|
||||
@ -274,9 +197,8 @@ pull_request_rules:
|
||||
|
||||
- name: ping author on conflicts and add 'needs-rebase' label
|
||||
conditions:
|
||||
- label != stale
|
||||
- conflict
|
||||
- -closed
|
||||
- conflict
|
||||
- -closed
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
@ -290,55 +212,20 @@ pull_request_rules:
|
||||
|
||||
- name: assign reviewer for tensorizer changes
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^vllm/model_executor/model_loader/tensorizer.py
|
||||
- files~=^vllm/model_executor/model_loader/tensorizer_loader.py
|
||||
- files~=^tests/entrypoints/openai/test_tensorizer_entrypoint.py
|
||||
- files~=^tests/model_executor/model_loader/tensorizer_loader/
|
||||
- files~=^tests/tensorizer_loader/
|
||||
actions:
|
||||
assign:
|
||||
users:
|
||||
- "sangstar"
|
||||
|
||||
- name: assign reviewer for modelopt changes
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^vllm/model_executor/layers/quantization/modelopt\.py$
|
||||
- files~=^vllm/model_executor/layers/quantization/__init__\.py$
|
||||
- files~=^tests/models/quantization/test_modelopt\.py$
|
||||
- files~=^tests/quantization/test_modelopt\.py$
|
||||
- files~=^tests/models/quantization/test_nvfp4\.py$
|
||||
- files~=^docs/features/quantization/modelopt\.md$
|
||||
actions:
|
||||
assign:
|
||||
users:
|
||||
- "Edwardf0t1"
|
||||
|
||||
- name: remove 'needs-rebase' label when conflict is resolved
|
||||
conditions:
|
||||
- -conflict
|
||||
- -closed
|
||||
- -conflict
|
||||
- -closed
|
||||
actions:
|
||||
label:
|
||||
remove:
|
||||
- needs-rebase
|
||||
|
||||
- name: label-kv-connector
|
||||
description: Automatically apply kv-connector label
|
||||
conditions:
|
||||
- label != stale
|
||||
- or:
|
||||
- files~=^examples/online_serving/disaggregated[^/]*/.*
|
||||
- files~=^examples/offline_inference/disaggregated[^/]*/.*
|
||||
- files~=^examples/others/lmcache/
|
||||
- files~=^tests/v1/kv_connector/
|
||||
- files~=^vllm/distributed/kv_transfer/
|
||||
- title~=(?i)\bP/?D\b
|
||||
- title~=(?i)NIXL
|
||||
- title~=(?i)LMCache
|
||||
actions:
|
||||
label:
|
||||
add:
|
||||
- kv-connector
|
||||
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}"
|
||||
|
||||
2
.github/workflows/add_label_automerge.yml
vendored
2
.github/workflows/add_label_automerge.yml
vendored
@ -10,7 +10,7 @@ jobs:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Add label
|
||||
uses: actions/github-script@ed597411d8f924073f98dfc5c65a23a2325f34cd # v8.0.0
|
||||
uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
|
||||
with:
|
||||
script: |
|
||||
github.rest.issues.addLabels({
|
||||
|
||||
29
.github/workflows/bc-lint.yml
vendored
29
.github/workflows/bc-lint.yml
vendored
@ -1,29 +0,0 @@
|
||||
name: BC Lint
|
||||
|
||||
on:
|
||||
pull_request:
|
||||
types:
|
||||
- opened
|
||||
- synchronize
|
||||
- reopened
|
||||
- labeled
|
||||
- unlabeled
|
||||
|
||||
jobs:
|
||||
bc_lint:
|
||||
if: github.repository_owner == 'vllm-project'
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Run BC Lint Action
|
||||
uses: pytorch/test-infra/.github/actions/bc-lint@main
|
||||
with:
|
||||
repo: ${{ github.event.pull_request.head.repo.full_name }}
|
||||
base_sha: ${{ github.event.pull_request.base.sha }}
|
||||
head_sha: ${{ github.event.pull_request.head.sha }}
|
||||
suppression: ${{ contains(github.event.pull_request.labels.*.name, 'suppress-bc-linter') }}
|
||||
docs_link: 'https://github.com/pytorch/test-infra/wiki/BC-Linter'
|
||||
config_dir: .github
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}
|
||||
cancel-in-progress: true
|
||||
2
.github/workflows/cleanup_pr_body.yml
vendored
2
.github/workflows/cleanup_pr_body.yml
vendored
@ -16,7 +16,7 @@ jobs:
|
||||
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
|
||||
- name: Set up Python
|
||||
uses: actions/setup-python@e797f83bcb11b83ae66e0230d6156d7c80228e7c # v6.0.0
|
||||
uses: actions/setup-python@42375524e23c412d93fb67b49958b491fce71c38 # v5.4.0
|
||||
with:
|
||||
python-version: '3.12'
|
||||
|
||||
|
||||
361
.github/workflows/issue_autolabel.yml
vendored
361
.github/workflows/issue_autolabel.yml
vendored
@ -1,361 +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
|
||||
id: label-step
|
||||
uses: actions/github-script@ed597411d8f924073f98dfc5c65a23a2325f34cd # v8.0.0
|
||||
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"
|
||||
}
|
||||
],
|
||||
},
|
||||
// Add more label configurations here as needed
|
||||
// example: {
|
||||
// keywords: [...],
|
||||
// substrings: [...],
|
||||
// regexPatterns: [...]
|
||||
// },
|
||||
};
|
||||
// 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 labelsAddedResults = await Promise.all(
|
||||
Object.entries(labelConfig).map(([labelName, config]) =>
|
||||
processLabel(labelName, config).then(added => ({ labelName, added }))
|
||||
)
|
||||
);
|
||||
|
||||
const numLabelsAdded = labelsAddedResults.filter(r => r.added).length;
|
||||
core.notice(`Processing complete. ${numLabelsAdded} label(s) added.`);
|
||||
|
||||
// Return which labels were added for the next step
|
||||
const addedLabels = labelsAddedResults.filter(r => r.added).map(r => r.labelName);
|
||||
core.setOutput('labels_added', JSON.stringify(addedLabels));
|
||||
return addedLabels;
|
||||
|
||||
- name: CC users for labeled issues
|
||||
if: steps.label-step.outputs.labels_added != '[]'
|
||||
uses: actions/github-script@ed597411d8f924073f98dfc5c65a23a2325f34cd # v8.0.0
|
||||
with:
|
||||
script: |
|
||||
// Configuration: Map labels to GitHub users to CC
|
||||
// You can add multiple users per label, and multiple label configurations
|
||||
const ccConfig = {
|
||||
rocm: {
|
||||
users: ['hongxiayang', 'tjtanaa', 'vllmellm'], // Add more users as needed: ['user1', 'user2', 'user3']
|
||||
message: 'CC {users} for ROCm-related issue' // {users} will be replaced with @mentions
|
||||
},
|
||||
// Add more label -> user mappings here
|
||||
// Example:
|
||||
// cuda: {
|
||||
// users: ['user1', 'user2'],
|
||||
// message: 'CC {users} for CUDA-related issue'
|
||||
// },
|
||||
// performance: {
|
||||
// users: ['perfexpert'],
|
||||
// message: 'CC {users} for performance issue'
|
||||
// },
|
||||
};
|
||||
|
||||
const labelsAdded = JSON.parse('${{ steps.label-step.outputs.labels_added }}');
|
||||
core.notice(`Labels added: ${labelsAdded.join(', ')}`);
|
||||
|
||||
// Get existing comments to check for already mentioned users
|
||||
const comments = await github.rest.issues.listComments({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
});
|
||||
|
||||
const issueBody = context.payload.issue.body || '';
|
||||
const allExistingText = issueBody + '\n' + comments.data.map(c => c.body).join('\n');
|
||||
|
||||
// Process each label that was added
|
||||
for (const label of labelsAdded) {
|
||||
if (ccConfig[label]) {
|
||||
const config = ccConfig[label];
|
||||
const usersToMention = [];
|
||||
|
||||
// Check which users haven't been mentioned yet
|
||||
for (const user of config.users) {
|
||||
const mentionPattern = new RegExp(`@${user}\\b`, 'i');
|
||||
if (!mentionPattern.test(allExistingText)) {
|
||||
usersToMention.push(user);
|
||||
} else {
|
||||
core.notice(`@${user} already mentioned for label "${label}", skipping`);
|
||||
}
|
||||
}
|
||||
|
||||
// Post comment if there are users to mention
|
||||
if (usersToMention.length > 0) {
|
||||
const mentions = usersToMention.map(u => `@${u}`).join(' ');
|
||||
const message = config.message.replace('{users}', mentions);
|
||||
|
||||
await github.rest.issues.createComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
body: message
|
||||
});
|
||||
|
||||
core.notice(`CC comment added for label "${label}": ${mentions}`);
|
||||
} else {
|
||||
core.notice(`All users for label "${label}" already mentioned, skipping comment`);
|
||||
}
|
||||
}
|
||||
}
|
||||
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
|
||||
}
|
||||
]
|
||||
}
|
||||
]
|
||||
}
|
||||
7
.github/workflows/pre-commit.yml
vendored
7
.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
|
||||
|
||||
@ -17,11 +13,10 @@ jobs:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
|
||||
- uses: actions/setup-python@e797f83bcb11b83ae66e0230d6156d7c80228e7c # v6.0.0
|
||||
- uses: actions/setup-python@42375524e23c412d93fb67b49958b491fce71c38 # v5.4.0
|
||||
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: ${{ secrets.PYPI_API_TOKEN }}
|
||||
# skip-existing: true
|
||||
51
.github/workflows/reminder_comment.yml
vendored
51
.github/workflows/reminder_comment.yml
vendored
@ -9,46 +9,19 @@ jobs:
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- name: Remind to run full CI on PR
|
||||
uses: actions/github-script@ed597411d8f924073f98dfc5c65a23a2325f34cd # v8.0.0
|
||||
uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
|
||||
with:
|
||||
script: |
|
||||
try {
|
||||
// Get the PR author
|
||||
const prAuthor = context.payload.pull_request.user.login;
|
||||
|
||||
// Check if this is the author's first PR in this repository
|
||||
// Use GitHub's search API to find all PRs by this author
|
||||
const { data: searchResults } = await github.rest.search.issuesAndPullRequests({
|
||||
q: `repo:${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
|
||||
|
||||
|
||||
2
.github/workflows/stale.yml
vendored
2
.github/workflows/stale.yml
vendored
@ -13,7 +13,7 @@ jobs:
|
||||
actions: write
|
||||
runs-on: ubuntu-latest
|
||||
steps:
|
||||
- uses: actions/stale@5f858e3efba33a5ca4407a664cc011ad407f2008 # v10.1.0
|
||||
- uses: actions/stale@5bef64f19d7facfb25b37b414482c7164d639639 # v9.1.0
|
||||
with:
|
||||
# Increasing this value ensures that changes to this workflow
|
||||
# propagate to all issues and PRs in days rather than months
|
||||
|
||||
18
.gitignore
vendored
18
.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/
|
||||
@ -177,14 +172,6 @@ cython_debug/
|
||||
# VSCode
|
||||
.vscode/
|
||||
|
||||
# Claude
|
||||
CLAUDE.md
|
||||
.claude/
|
||||
|
||||
# Codex
|
||||
AGENTS.md
|
||||
.codex/
|
||||
|
||||
# DS Store
|
||||
.DS_Store
|
||||
|
||||
@ -215,6 +202,3 @@ shellcheck*/
|
||||
|
||||
# Ignore moe/marlin_moe gen code
|
||||
csrc/moe/marlin_moe_wna16/kernel_*
|
||||
|
||||
# Ignore ep_kernels_workspace folder
|
||||
ep_kernels_workspace/
|
||||
|
||||
@ -1,12 +0,0 @@
|
||||
MD007:
|
||||
indent: 4
|
||||
MD013: false
|
||||
MD024:
|
||||
siblings_only: true
|
||||
MD033: false
|
||||
MD045: false
|
||||
MD046: false
|
||||
MD051: false
|
||||
MD052: false
|
||||
MD053: false
|
||||
MD059: false
|
||||
@ -6,75 +6,87 @@ default_stages:
|
||||
- manual # Run in CI
|
||||
exclude: 'vllm/third_party/.*'
|
||||
repos:
|
||||
- repo: https://github.com/astral-sh/ruff-pre-commit
|
||||
rev: v0.14.0
|
||||
- repo: https://github.com/google/yapf
|
||||
rev: v0.43.0
|
||||
hooks:
|
||||
- id: ruff-check
|
||||
- id: yapf
|
||||
args: [--in-place, --verbose]
|
||||
# Keep the same list from yapfignore here to avoid yapf failing without any inputs
|
||||
exclude: '(.buildkite|benchmarks|build|examples)/.*'
|
||||
- repo: https://github.com/astral-sh/ruff-pre-commit
|
||||
rev: v0.11.7
|
||||
hooks:
|
||||
- id: ruff
|
||||
args: [--output-format, github, --fix]
|
||||
- id: ruff-format
|
||||
files: ^(.buildkite|benchmarks|examples)/.*
|
||||
- repo: https://github.com/crate-ci/typos
|
||||
rev: v1.38.1
|
||||
rev: v1.32.0
|
||||
hooks:
|
||||
- id: typos
|
||||
args: [--force-exclude]
|
||||
- repo: https://github.com/PyCQA/isort
|
||||
rev: 6.0.1
|
||||
hooks:
|
||||
- id: isort
|
||||
- repo: https://github.com/pre-commit/mirrors-clang-format
|
||||
rev: v21.1.2
|
||||
rev: v20.1.3
|
||||
hooks:
|
||||
- id: clang-format
|
||||
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:
|
||||
- id: actionlint
|
||||
- repo: https://github.com/astral-sh/uv-pre-commit
|
||||
rev: 0.9.1
|
||||
rev: 0.6.17
|
||||
hooks:
|
||||
- id: pip-compile
|
||||
args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match, --torch-backend, cu128, --python-platform, x86_64-manylinux_2_28]
|
||||
args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match, --torch-backend, cu128]
|
||||
files: ^requirements/test\.(in|txt)$
|
||||
- repo: local
|
||||
hooks:
|
||||
- id: format-torch-nightly-test
|
||||
name: reformat nightly_torch_test.txt to be in sync with test.in
|
||||
language: python
|
||||
entry: python tools/generate_nightly_torch_test.py
|
||||
files: ^requirements/test\.(in|txt)$
|
||||
- id: mypy-local
|
||||
name: Run mypy for local Python installation
|
||||
entry: python tools/pre_commit/mypy.py 0 "local"
|
||||
entry: tools/mypy.sh 0 "local"
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: &mypy_deps [mypy==1.11.1, types-cachetools, types-setuptools, types-PyYAML, types-requests, pydantic]
|
||||
stages: [pre-commit] # Don't run in CI
|
||||
<<: &mypy_common
|
||||
language: python
|
||||
types_or: [python, pyi]
|
||||
require_serial: true
|
||||
additional_dependencies: [mypy==1.11.1, regex, types-cachetools, types-setuptools, types-PyYAML, types-requests, types-torch, pydantic]
|
||||
- id: mypy-3.9 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
|
||||
name: Run mypy for Python 3.9
|
||||
entry: tools/mypy.sh 1 "3.9"
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: *mypy_deps
|
||||
stages: [manual] # Only run in CI
|
||||
- id: mypy-3.10 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
|
||||
name: Run mypy for Python 3.10
|
||||
entry: python tools/pre_commit/mypy.py 1 "3.10"
|
||||
<<: *mypy_common
|
||||
entry: tools/mypy.sh 1 "3.10"
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: *mypy_deps
|
||||
stages: [manual] # Only run in CI
|
||||
- id: mypy-3.11 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
|
||||
name: Run mypy for Python 3.11
|
||||
entry: python tools/pre_commit/mypy.py 1 "3.11"
|
||||
<<: *mypy_common
|
||||
entry: tools/mypy.sh 1 "3.11"
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: *mypy_deps
|
||||
stages: [manual] # Only run in CI
|
||||
- id: mypy-3.12 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
|
||||
name: Run mypy for Python 3.12
|
||||
entry: python tools/pre_commit/mypy.py 1 "3.12"
|
||||
<<: *mypy_common
|
||||
stages: [manual] # Only run in CI
|
||||
- id: mypy-3.13 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
|
||||
name: Run mypy for Python 3.13
|
||||
entry: python tools/pre_commit/mypy.py 1 "3.13"
|
||||
<<: *mypy_common
|
||||
entry: tools/mypy.sh 1 "3.12"
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: *mypy_deps
|
||||
stages: [manual] # Only run in CI
|
||||
- id: shellcheck
|
||||
name: Lint shell scripts
|
||||
@ -103,11 +115,6 @@ repos:
|
||||
entry: python tools/check_spdx_header.py
|
||||
language: python
|
||||
types: [python]
|
||||
- id: check-root-lazy-imports
|
||||
name: Check root lazy imports
|
||||
entry: python tools/check_init_lazy_imports.py
|
||||
language: python
|
||||
types: [python]
|
||||
- id: check-filenames
|
||||
name: Check for spaces in all filenames
|
||||
entry: bash
|
||||
@ -138,19 +145,15 @@ repos:
|
||||
additional_dependencies: [regex]
|
||||
- id: check-pickle-imports
|
||||
name: Prevent new pickle/cloudpickle imports
|
||||
entry: python tools/pre_commit/check_pickle_imports.py
|
||||
entry: python tools/check_pickle_imports.py
|
||||
language: python
|
||||
types: [python]
|
||||
additional_dependencies: [regex]
|
||||
- id: validate-config
|
||||
name: Validate configuration has default values and that each field has a docstring
|
||||
entry: python tools/validate_config.py
|
||||
language: python
|
||||
additional_dependencies: [regex]
|
||||
pass_filenames: false
|
||||
additional_dependencies: [pathspec, regex]
|
||||
# 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
|
||||
|
||||
@ -7,13 +7,9 @@ build:
|
||||
os: ubuntu-22.04
|
||||
tools:
|
||||
python: "3.12"
|
||||
jobs:
|
||||
post_checkout:
|
||||
- git fetch --unshallow || true
|
||||
|
||||
mkdocs:
|
||||
configuration: mkdocs.yaml
|
||||
fail_on_warning: true
|
||||
|
||||
# Optionally declare the Python requirements required to build your docs
|
||||
python:
|
||||
|
||||
@ -1,2 +1 @@
|
||||
collect_env.py
|
||||
vllm/model_executor/layers/fla/ops/*.py
|
||||
|
||||
329
CMakeLists.txt
329
CMakeLists.txt
@ -13,10 +13,6 @@ cmake_minimum_required(VERSION 3.26)
|
||||
# cmake --install . --component _C
|
||||
project(vllm_extensions LANGUAGES CXX)
|
||||
|
||||
set(CMAKE_CXX_STANDARD 17)
|
||||
set(CMAKE_CXX_STANDARD_REQUIRED ON)
|
||||
|
||||
|
||||
# CUDA by default, can be overridden by using -DVLLM_TARGET_DEVICE=... (used by setup.py)
|
||||
set(VLLM_TARGET_DEVICE "cuda" CACHE STRING "Target device backend for vLLM")
|
||||
message(STATUS "Build type: ${CMAKE_BUILD_TYPE}")
|
||||
@ -34,10 +30,10 @@ 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 "3.10" "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;gfx1150;gfx1151")
|
||||
set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1030;gfx1100;gfx1101;gfx1200;gfx1201")
|
||||
|
||||
#
|
||||
# Supported/expected torch versions for CUDA/ROCm.
|
||||
@ -49,8 +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
|
||||
@ -86,9 +82,6 @@ find_package(Torch REQUIRED)
|
||||
# Supported NVIDIA architectures.
|
||||
# This check must happen after find_package(Torch) because that's when CMAKE_CUDA_COMPILER_VERSION gets defined
|
||||
if(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
|
||||
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
|
||||
set(CUDA_SUPPORTED_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;11.0;12.0")
|
||||
elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
|
||||
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.8)
|
||||
set(CUDA_SUPPORTED_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;12.0")
|
||||
else()
|
||||
@ -178,24 +171,6 @@ if(NVCC_THREADS AND VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND VLLM_GPU_FLAGS "--threads=${NVCC_THREADS}")
|
||||
endif()
|
||||
|
||||
#
|
||||
# Set compression mode for CUDA >=13.x.
|
||||
#
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA" AND
|
||||
DEFINED CMAKE_CUDA_COMPILER_VERSION AND
|
||||
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
|
||||
list(APPEND VLLM_GPU_FLAGS "--compress-mode=size")
|
||||
endif()
|
||||
|
||||
#
|
||||
# Set CUDA include flags for CXX compiler.
|
||||
#
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -I${CUDA_TOOLKIT_ROOT_DIR}/include")
|
||||
if(CUDA_VERSION VERSION_GREATER_EQUAL 13.0)
|
||||
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -I${CUDA_TOOLKIT_ROOT_DIR}/include/cccl")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
#
|
||||
# Use FetchContent for C++ dependencies that are compiled as part of vLLM's build process.
|
||||
@ -257,6 +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"
|
||||
@ -269,12 +245,13 @@ set(VLLM_EXT_SRC
|
||||
"csrc/sampler.cu"
|
||||
"csrc/cuda_view.cu"
|
||||
"csrc/quantization/gptq/q_gemm.cu"
|
||||
"csrc/quantization/w8a8/int8/scaled_quant.cu"
|
||||
"csrc/quantization/w8a8/fp8/common.cu"
|
||||
"csrc/quantization/compressed_tensors/int8_quant_kernels.cu"
|
||||
"csrc/quantization/fp8/common.cu"
|
||||
"csrc/quantization/fused_kernels/fused_layernorm_dynamic_per_token_quant.cu"
|
||||
"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")
|
||||
|
||||
@ -282,7 +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 "v4.2.1" 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})
|
||||
@ -312,15 +289,16 @@ 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/w8a8/cutlass/scaled_mm_entry.cu"
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_entry.cu"
|
||||
"csrc/quantization/fp4/nvfp4_quant_entry.cu"
|
||||
"csrc/quantization/fp4/nvfp4_scaled_mm_entry.cu"
|
||||
"csrc/quantization/fp4/nvfp4_blockwise_moe_kernel.cu"
|
||||
"csrc/sparse/cutlass/sparse_scaled_mm_entry.cu"
|
||||
"csrc/cutlass_extensions/common.cpp"
|
||||
"csrc/quantization/w8a8/fp8/per_token_group_quant.cu"
|
||||
"csrc/quantization/w8a8/int8/per_token_group_quant.cu")
|
||||
"csrc/attention/mla/cutlass_mla_entry.cu")
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${VLLM_EXT_SRC}"
|
||||
@ -374,27 +352,20 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${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: ${MARLIN_ARCHS}")
|
||||
else()
|
||||
message(STATUS "Not building Marlin kernels as no compatible archs found"
|
||||
@ -422,13 +393,13 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The cutlass_scaled_mm kernels for Hopper (c3x, i.e. CUTLASS 3.x) require
|
||||
# CUDA 12.0 or later
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a;" "${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/w8a8/cutlass/scaled_mm_c3x_sm90.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_sm90_fp8.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_sm90_int8.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_azp_sm90_int8.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_blockwise_sm90_fp8.cu")
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm90.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm90_fp8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm90_int8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_azp_sm90_int8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_blockwise_sm90_fp8.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
@ -438,7 +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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building scaled_mm_c3x_sm90 as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
"later if you intend on running FP8 quantized models on "
|
||||
@ -449,53 +420,14 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
|
||||
# The cutlass_scaled_mm kernels for Geforce Blackwell SM120 (c3x, i.e. CUTLASS 3.x) require
|
||||
# CUDA 12.8 or later
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/w8a8/cutlass/scaled_mm_c3x_sm120.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_sm120_fp8.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_blockwise_sm120_fp8.cu"
|
||||
)
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_SCALED_MM_SM120=1")
|
||||
# Let scaled_mm_c2x know it doesn't need to build these arches
|
||||
list(APPEND SCALED_MM_3X_ARCHS "${SCALED_MM_ARCHS}")
|
||||
message(STATUS "Building scaled_mm_c3x_sm120 for archs: ${SCALED_MM_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building scaled_mm_c3x_sm120 as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or "
|
||||
"later if you intend on running FP8 quantized models on "
|
||||
"Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building scaled_mm_c3x_120 as no compatible archs found "
|
||||
"in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
|
||||
# The cutlass_scaled_mm kernels for Blackwell SM100 (c3x, i.e. CUTLASS 3.x)
|
||||
# require CUDA 12.8 or later
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/w8a8/cutlass/scaled_mm_c3x_sm100.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_sm100_fp8.cu"
|
||||
"csrc/quantization/w8a8/cutlass/c3x/scaled_mm_blockwise_sm100_fp8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm100.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_sm100_fp8.cu"
|
||||
"csrc/quantization/cutlass_w8a8/c3x/scaled_mm_blockwise_sm100_fp8.cu"
|
||||
)
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
@ -506,7 +438,7 @@ 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 "
|
||||
@ -526,7 +458,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# subtract out the archs that are already built for 3x
|
||||
list(REMOVE_ITEM SCALED_MM_2X_ARCHS ${SCALED_MM_3X_ARCHS})
|
||||
if (SCALED_MM_2X_ARCHS)
|
||||
set(SRCS "csrc/quantization/w8a8/cutlass/scaled_mm_c2x.cu")
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/scaled_mm_c2x.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_2X_ARCHS}")
|
||||
@ -549,7 +481,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The 2:4 sparse kernels cutlass_scaled_sparse_mm and cutlass_compressor
|
||||
# require CUDA 12.2 or later (and only work on Hopper).
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a;" "${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}"
|
||||
@ -558,7 +490,7 @@ 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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.2 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building sparse_scaled_mm_c3x kernels as CUDA Compiler version is "
|
||||
"not >= 12.2, we recommend upgrading to CUDA 12.2 or later "
|
||||
"if you intend on running FP8 sparse quantized models on Hopper.")
|
||||
@ -568,40 +500,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# The nvfp4_scaled_mm_sm120 kernels for Geforce Blackwell SM120 require
|
||||
# CUDA 12.8 or later
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "12.0a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
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: ${FP4_ARCHS}")
|
||||
else()
|
||||
message(STATUS "Not building NVFP4 as no compatible archs were found.")
|
||||
# clear FP4_ARCHS
|
||||
set(FP4_ARCHS)
|
||||
endif()
|
||||
|
||||
# FP4 Archs and flags
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "10.0a;10.1a;12.0a;12.1a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND FP4_ARCHS)
|
||||
cuda_archs_loose_intersection(FP4_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND FP4_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/quantization/fp4/nvfp4_quant_kernels.cu"
|
||||
"csrc/quantization/fp4/activation_nvfp4_quant_fusion_kernels.cu"
|
||||
"csrc/quantization/fp4/nvfp4_experts_quant.cu"
|
||||
"csrc/quantization/fp4/nvfp4_scaled_mm_kernels.cu"
|
||||
"csrc/quantization/fp4/nvfp4_blockwise_moe_kernel.cu")
|
||||
@ -609,8 +512,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
SRCS "${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_CUTLASS_MOE_SM100=1")
|
||||
list(APPEND VLLM_GPU_FLAGS "-DENABLE_NVFP4=1")
|
||||
message(STATUS "Building NVFP4 for archs: ${FP4_ARCHS}")
|
||||
else()
|
||||
message(STATUS "Not building NVFP4 as no compatible archs were found.")
|
||||
@ -619,14 +521,10 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
|
||||
# CUTLASS MLA Archs and flags
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(MLA_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(MLA_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND MLA_ARCHS)
|
||||
cuda_archs_loose_intersection(MLA_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8 AND MLA_ARCHS)
|
||||
set(SRCS
|
||||
"csrc/attention/mla/sm100_cutlass_mla_kernel.cu")
|
||||
"csrc/attention/mla/cutlass_mla_kernels.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${MLA_ARCHS}")
|
||||
@ -649,7 +547,8 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# if it's possible to compile MoE kernels that use its output.
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "9.0a" "${CUDA_ARCHS}")
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/w8a8/cutlass/moe/grouped_mm_c3x_sm90.cu")
|
||||
set(SRCS "csrc/quantization/cutlass_w8a8/moe/grouped_mm_c3x.cu"
|
||||
"csrc/quantization/cutlass_w8a8/moe/moe_data.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${SCALED_MM_ARCHS}")
|
||||
@ -663,78 +562,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
"if you intend on running FP8 quantized MoE models on Hopper.")
|
||||
else()
|
||||
message(STATUS "Not building grouped_mm_c3x as no compatible archs found "
|
||||
"in CUDA target architectures.")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/w8a8/cutlass/moe/grouped_mm_c3x_sm100.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building grouped_mm_c3x kernels as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or later "
|
||||
"if you intend on running FP8 quantized MoE models on Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building grouped_mm_c3x as no compatible archs found "
|
||||
"in CUDA target architectures.")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# moe_data.cu is used by all CUTLASS MoE kernels.
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND CUTLASS_MOE_DATA_ARCHS)
|
||||
set(SRCS "csrc/quantization/w8a8/cutlass/moe/moe_data.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${CUTLASS_MOE_DATA_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
message(STATUS "Building moe_data for archs: ${CUTLASS_MOE_DATA_ARCHS}")
|
||||
else()
|
||||
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND CUTLASS_MOE_DATA_ARCHS)
|
||||
message(STATUS "Not building moe_data as CUDA Compiler version is "
|
||||
"not >= 12.3, we recommend upgrading to CUDA 12.3 or later "
|
||||
"if you intend on running FP8 quantized MoE models on Hopper or Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building moe_data as no compatible archs found "
|
||||
"in CUDA target architectures.")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
|
||||
else()
|
||||
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
|
||||
endif()
|
||||
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
set(SRCS "csrc/quantization/w8a8/cutlass/moe/blockwise_scaled_group_mm_sm100.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
|
||||
message(STATUS "Not building blockwise_scaled_group_mm_sm100 kernels as CUDA Compiler version is "
|
||||
"not >= 12.8, we recommend upgrading to CUDA 12.8 or later "
|
||||
"if you intend on running FP8 quantized MoE models on Blackwell.")
|
||||
else()
|
||||
message(STATUS "Not building blockwise_scaled_group_mm_sm100 as no compatible archs found "
|
||||
"in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
@ -745,7 +572,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
# The machete kernels only work on hopper and require CUDA 12.0 or later.
|
||||
# Only build Machete kernels if we are building for something compatible with sm90a
|
||||
cuda_archs_loose_intersection(MACHETE_ARCHS "9.0a" "${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.
|
||||
@ -797,7 +624,7 @@ 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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.0
|
||||
AND MACHETE_ARCHS)
|
||||
message(STATUS "Not building Machete kernels as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
@ -808,55 +635,9 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
"found in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# Only build W4A8 kernels if we are building for something compatible with sm90a
|
||||
cuda_archs_loose_intersection(W4A8_ARCHS "9.0a" "${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 ${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0
|
||||
AND W4A8_ARCHS)
|
||||
message(STATUS "Not building W4A8 kernels as CUDA Compiler version is "
|
||||
"not >= 12.0, we recommend upgrading to CUDA 12.0 or "
|
||||
"later if you intend on running w4a16 quantized models on "
|
||||
"Hopper.")
|
||||
else()
|
||||
message(STATUS "Not building W4A8 kernels as no compatible archs "
|
||||
"found in CUDA target architectures")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# Hadacore kernels
|
||||
cuda_archs_loose_intersection(HADACORE_ARCHS "8.0;8.9;9.0" "${CUDA_ARCHS}")
|
||||
if(HADACORE_ARCHS)
|
||||
set(SRCS "csrc/quantization/hadamard/hadacore/hadamard_transform_cuda.cu")
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${SRCS}"
|
||||
CUDA_ARCHS "${HADACORE_ARCHS}")
|
||||
list(APPEND VLLM_EXT_SRC "${SRCS}")
|
||||
message(STATUS "Building hadacore")
|
||||
endif()
|
||||
|
||||
# if CUDA endif
|
||||
endif()
|
||||
|
||||
if (VLLM_GPU_LANG STREQUAL "HIP")
|
||||
# Add QuickReduce kernels
|
||||
list(APPEND VLLM_EXT_SRC
|
||||
"csrc/custom_quickreduce.cu"
|
||||
)
|
||||
# if ROCM endif
|
||||
endif()
|
||||
|
||||
message(STATUS "Enabling C extension.")
|
||||
define_gpu_extension_target(
|
||||
_C
|
||||
@ -886,17 +667,7 @@ set(VLLM_MOE_EXT_SRC
|
||||
"csrc/moe/topk_softmax_kernels.cu")
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
list(APPEND VLLM_MOE_EXT_SRC
|
||||
"csrc/moe/moe_wna16.cu"
|
||||
"csrc/moe/grouped_topk_kernels.cu")
|
||||
endif()
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set(MOE_PERMUTE_SRC
|
||||
"csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
|
||||
"csrc/moe/moe_permute_unpermute_op.cu")
|
||||
|
||||
list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
|
||||
list(APPEND VLLM_MOE_EXT_SRC "csrc/moe/moe_wna16.cu")
|
||||
endif()
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
@ -957,10 +728,6 @@ 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 ${MOE_WNAA16_MARLIN_SRC})
|
||||
|
||||
@ -971,6 +738,17 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
if(VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
set(MOE_PERMUTE_SRC
|
||||
"csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
|
||||
"csrc/moe/moe_permute_unpermute_op.cu")
|
||||
|
||||
set_gencode_flags_for_srcs(
|
||||
SRCS "${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
|
||||
@ -1007,7 +785,6 @@ endif()
|
||||
# For CUDA we also build and ship some external projects.
|
||||
if (VLLM_GPU_LANG STREQUAL "CUDA")
|
||||
include(cmake/external_projects/flashmla.cmake)
|
||||
include(cmake/external_projects/qutlass.cmake)
|
||||
|
||||
# vllm-flash-attn should be last as it overwrites some CMake functions
|
||||
include(cmake/external_projects/vllm_flash_attn.cmake)
|
||||
|
||||
@ -2,6 +2,7 @@ include LICENSE
|
||||
include requirements/common.txt
|
||||
include requirements/cuda.txt
|
||||
include requirements/rocm.txt
|
||||
include requirements/neuron.txt
|
||||
include requirements/cpu.txt
|
||||
include CMakeLists.txt
|
||||
|
||||
|
||||
32
README.md
32
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">
|
||||
@ -14,27 +13,17 @@ Easy, fast, and cheap LLM serving for everyone
|
||||
| <a href="https://docs.vllm.ai"><b>Documentation</b></a> | <a href="https://blog.vllm.ai/"><b>Blog</b></a> | <a href="https://arxiv.org/abs/2309.06180"><b>Paper</b></a> | <a href="https://x.com/vllm_project"><b>Twitter/X</b></a> | <a href="https://discuss.vllm.ai"><b>User Forum</b></a> | <a href="https://slack.vllm.ai"><b>Developer Slack</b></a> |
|
||||
</p>
|
||||
|
||||
---
|
||||
Join us at the [PyTorch Conference, October 22-23](https://events.linuxfoundation.org/pytorch-conference/) and [Ray Summit, November 3-5](https://www.anyscale.com/ray-summit/2025) in San Francisco for our latest updates on vLLM and to meet the vLLM team! Register now for the largest vLLM community events of the year!
|
||||
|
||||
---
|
||||
|
||||
*Latest News* 🔥
|
||||
|
||||
- [2025/09] We hosted [vLLM Toronto Meetup](https://luma.com/e80e0ymm) focused on tackling inference at scale and speculative decoding with speakers from NVIDIA and Red Hat! Please find the meetup slides [here](https://docs.google.com/presentation/d/1IYJYmJcu9fLpID5N5RbW_vO0XLo0CGOR14IXOjB61V8/edit?usp=sharing).
|
||||
- [2025/08] We hosted [vLLM Shenzhen Meetup](https://mp.weixin.qq.com/s/k8ZBO1u2_2odgiKWH_GVTQ) focusing on the ecosystem around vLLM! Please find the meetup slides [here](https://drive.google.com/drive/folders/1Ua2SVKVSu-wp5vou_6ElraDt2bnKhiEA).
|
||||
- [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).
|
||||
- [2025/05] 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>
|
||||
|
||||
- [2025/08] 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).
|
||||
- [2025/04] 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).
|
||||
@ -57,7 +46,6 @@ Join us at the [PyTorch Conference, October 22-23](https://events.linuxfoundatio
|
||||
</details>
|
||||
|
||||
---
|
||||
|
||||
## About
|
||||
|
||||
vLLM is a fast and easy-to-use library for LLM inference and serving.
|
||||
@ -75,19 +63,20 @@ 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 for NVIDIA GPUs, AMD CPUs and GPUs, Intel CPUs and GPUs, PowerPC CPUs, and TPU. Additionally, support for diverse hardware plugins such as Intel Gaudi, IBM Spyre and Huawei Ascend.
|
||||
- Support NVIDIA GPUs, AMD CPUs and GPUs, Intel CPUs and GPUs, PowerPC CPUs, TPU, and AWS Neuron
|
||||
- Prefix caching support
|
||||
- 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)
|
||||
@ -104,7 +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)
|
||||
@ -121,7 +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
|
||||
@ -129,8 +116,6 @@ Cash Donations:
|
||||
- ZhenFund
|
||||
|
||||
Compute Resources:
|
||||
|
||||
- Alibaba Cloud
|
||||
- AMD
|
||||
- Anyscale
|
||||
- AWS
|
||||
@ -149,7 +134,6 @@ Compute Resources:
|
||||
- Trainy
|
||||
- UC Berkeley
|
||||
- UC San Diego
|
||||
- Volcengine
|
||||
|
||||
Slack Sponsor: Anyscale
|
||||
|
||||
@ -170,13 +154,11 @@ 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
|
||||
- For collaborations and partnerships, please contact us at [vllm-questions@lists.berkeley.edu](mailto:vllm-questions@lists.berkeley.edu)
|
||||
<!-- --8<-- [end:contact-us] -->
|
||||
|
||||
## Media Kit
|
||||
|
||||
|
||||
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.
|
||||
|
||||
@ -1,20 +1,389 @@
|
||||
# Benchmarks
|
||||
# Benchmarking vLLM
|
||||
|
||||
This directory used to contain vLLM's benchmark scripts and utilities for performance testing and evaluation.
|
||||
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.
|
||||
|
||||
## Contents
|
||||
## Dataset Overview
|
||||
|
||||
- **Serving benchmarks**: Scripts for testing online inference performance (latency, throughput)
|
||||
- **Throughput benchmarks**: Scripts for testing offline batch inference performance
|
||||
- **Specialized benchmarks**: Tools for testing specific features like structured output, prefix caching, long document QA, request prioritization, and multi-modal inference
|
||||
- **Dataset utilities**: Framework for loading and sampling from various benchmark datasets (ShareGPT, HuggingFace datasets, synthetic data, etc.)
|
||||
<table style="width:100%; border-collapse: collapse;">
|
||||
<thead>
|
||||
<tr>
|
||||
<th style="width:15%; text-align: left;">Dataset</th>
|
||||
<th style="width:10%; text-align: center;">Online</th>
|
||||
<th style="width:10%; text-align: center;">Offline</th>
|
||||
<th style="width:65%; text-align: left;">Data Path</th>
|
||||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr>
|
||||
<td><strong>ShareGPT</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<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>BurstGPT</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>wget https://github.com/HPMLL/BurstGPT/releases/download/v1.1/BurstGPT_without_fails_2.csv</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<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>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>Random</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>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>lmarena-ai/VisionArena-Chat</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>HuggingFace-InstructCoder</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>likaixin/InstructCoder</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>HuggingFace-AIMO</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>AI-MO/aimo-validation-aime</code> , <code>AI-MO/NuminaMath-1.5</code>, <code>AI-MO/NuminaMath-CoT</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>HuggingFace-Other</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td><code>lmms-lab/LLaVA-OneVision-Data</code>, <code>Aeala/ShareGPT_Vicuna_unfiltered</code></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><strong>Custom</strong></td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td style="text-align: center;">✅</td>
|
||||
<td>Local file: <code>data.jsonl</code></td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
|
||||
## Usage
|
||||
✅: supported
|
||||
|
||||
For detailed usage instructions, examples, and dataset information, see the [Benchmark CLI documentation](https://docs.vllm.ai/en/latest/contributing/benchmarks.html#benchmark-cli).
|
||||
🟡: Partial support
|
||||
|
||||
For full CLI reference see:
|
||||
🚧: to be supported
|
||||
|
||||
- <https://docs.vllm.ai/en/latest/cli/bench/latency.html>
|
||||
- <https://docs.vllm.ai/en/latest/cli/bench/serve.html>
|
||||
- <https://docs.vllm.ai/en/latest/cli/bench/throughput.html>
|
||||
**Note**: HuggingFace dataset's `dataset-name` should be set to `hf`
|
||||
|
||||
---
|
||||
## Example - Online Benchmark
|
||||
|
||||
First start serving your model
|
||||
|
||||
```bash
|
||||
vllm serve NousResearch/Hermes-3-Llama-3.1-8B --disable-log-requests
|
||||
```
|
||||
|
||||
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
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--endpoint /v1/completions \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path <your data path>/ShareGPT_V3_unfiltered_cleaned_split.json \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
If successful, you will see the following output
|
||||
|
||||
```
|
||||
============ 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
|
||||
---------------Time to First Token----------------
|
||||
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
|
||||
---------------Inter-token Latency----------------
|
||||
Mean ITL (ms): 7.74
|
||||
Median ITL (ms): 7.70
|
||||
P99 ITL (ms): 8.39
|
||||
==================================================
|
||||
```
|
||||
|
||||
### 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
|
||||
|
||||
```
|
||||
{"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 --disable-log-requests
|
||||
```
|
||||
|
||||
```bash
|
||||
# run benchmarking script
|
||||
python3 benchmarks/benchmark_serving.py --port 9001 --save-result --save-detailed \
|
||||
--backend vllm \
|
||||
--model meta-llama/Llama-3.1-8B-Instruct \
|
||||
--endpoint /v1/completions \
|
||||
--dataset-name custom \
|
||||
--dataset-path <path-to-your-data-jsonl> \
|
||||
--custom-skip-chat-template \
|
||||
--num-prompts 80 \
|
||||
--max-concurrency 1 \
|
||||
--temperature=0.3 \
|
||||
--top-p=0.75 \
|
||||
--result-dir "./log/"
|
||||
```
|
||||
|
||||
You can skip applying chat template if your data already has it by using `--custom-skip-chat-template`.
|
||||
|
||||
### VisionArena Benchmark for Vision Language Models
|
||||
|
||||
```bash
|
||||
# need a model with vision capability here
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct --disable-log-requests
|
||||
```
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
--dataset-path lmarena-ai/VisionArena-Chat \
|
||||
--hf-split train \
|
||||
--num-prompts 1000
|
||||
```
|
||||
|
||||
### InstructCoder Benchmark with Speculative Decoding
|
||||
|
||||
``` bash
|
||||
VLLM_USE_V1=1 vllm serve meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
--speculative-config $'{"method": "ngram",
|
||||
"num_speculative_tokens": 5, "prompt_lookup_max": 5,
|
||||
"prompt_lookup_min": 2}'
|
||||
```
|
||||
|
||||
``` bash
|
||||
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
|
||||
|
||||
```bash
|
||||
vllm serve Qwen/Qwen2-VL-7B-Instruct --disable-log-requests
|
||||
```
|
||||
|
||||
**`lmms-lab/LLaVA-OneVision-Data`**
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
--dataset-path lmms-lab/LLaVA-OneVision-Data \
|
||||
--hf-split train \
|
||||
--hf-subset "chart2text(cauldron)" \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
**`Aeala/ShareGPT_Vicuna_unfiltered`**
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend openai-chat \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--endpoint /v1/chat/completions \
|
||||
--dataset-name hf \
|
||||
--dataset-path Aeala/ShareGPT_Vicuna_unfiltered \
|
||||
--hf-split train \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
**`AI-MO/aimo-validation-aime`**
|
||||
|
||||
``` bash
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--model Qwen/QwQ-32B \
|
||||
--dataset-name hf \
|
||||
--dataset-path AI-MO/aimo-validation-aime \
|
||||
--num-prompts 10 \
|
||||
--seed 42
|
||||
```
|
||||
|
||||
**`philschmid/mt-bench`**
|
||||
|
||||
``` bash
|
||||
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
|
||||
|
||||
When using OpenAI-compatible backends such as `vllm`, optional sampling
|
||||
parameters can be specified. Example client command:
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_serving.py \
|
||||
--backend vllm \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--endpoint /v1/completions \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path <your data path>/ShareGPT_V3_unfiltered_cleaned_split.json \
|
||||
--top-k 10 \
|
||||
--top-p 0.9 \
|
||||
--temperature 0.5 \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
---
|
||||
## Example - Offline Throughput Benchmark
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model NousResearch/Hermes-3-Llama-3.1-8B \
|
||||
--dataset-name sonnet \
|
||||
--dataset-path vllm/benchmarks/sonnet.txt \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
If successful, you will see the following output
|
||||
|
||||
```
|
||||
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
|
||||
|
||||
``` bash
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
--dataset-path lmarena-ai/VisionArena-Chat \
|
||||
--num-prompts 1000 \
|
||||
--hf-split train
|
||||
```
|
||||
|
||||
The `num prompt tokens` now includes image token counts
|
||||
|
||||
```
|
||||
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
|
||||
|
||||
``` bash
|
||||
VLLM_WORKER_MULTIPROC_METHOD=spawn \
|
||||
VLLM_USE_V1=1 \
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--dataset-name=hf \
|
||||
--dataset-path=likaixin/InstructCoder \
|
||||
--model=meta-llama/Meta-Llama-3-8B-Instruct \
|
||||
--input-len=1000 \
|
||||
--output-len=100 \
|
||||
--num-prompts=2048 \
|
||||
--async-engine \
|
||||
--speculative-config $'{"method": "ngram",
|
||||
"num_speculative_tokens": 5, "prompt_lookup_max": 5,
|
||||
"prompt_lookup_min": 2}'
|
||||
```
|
||||
|
||||
```
|
||||
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
|
||||
|
||||
**`lmms-lab/LLaVA-OneVision-Data`**
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
--dataset-path lmms-lab/LLaVA-OneVision-Data \
|
||||
--hf-split train \
|
||||
--hf-subset "chart2text(cauldron)" \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
**`Aeala/ShareGPT_Vicuna_unfiltered`**
|
||||
|
||||
```bash
|
||||
python3 vllm/benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/Qwen2-VL-7B-Instruct \
|
||||
--backend vllm-chat \
|
||||
--dataset-name hf \
|
||||
--dataset-path Aeala/ShareGPT_Vicuna_unfiltered \
|
||||
--hf-split train \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
**`AI-MO/aimo-validation-aime`**
|
||||
|
||||
```bash
|
||||
python3 benchmarks/benchmark_throughput.py \
|
||||
--model Qwen/QwQ-32B \
|
||||
--backend vllm \
|
||||
--dataset-name hf \
|
||||
--dataset-path AI-MO/aimo-validation-aime \
|
||||
--hf-split train \
|
||||
--num-prompts 10
|
||||
```
|
||||
|
||||
### Benchmark with LoRA Adapters
|
||||
|
||||
``` bash
|
||||
# download dataset
|
||||
# wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
|
||||
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 \
|
||||
--dataset_name sharegpt \
|
||||
--num-prompts 10 \
|
||||
--max-loras 2 \
|
||||
--max-lora-rank 8 \
|
||||
--enable-lora \
|
||||
--lora-path yard1/llama-2-7b-sql-lora-test
|
||||
```
|
||||
|
||||
@ -1,50 +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
|
||||
# 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=${NUM_BATCHED_TOKENS_LIST:-"512 1024 2048 4096"}
|
||||
BASE=""
|
||||
MODEL="meta-llama/Llama-3.1-8B-Instruct"
|
||||
TP=1
|
||||
DOWNLOAD_DIR=""
|
||||
INPUT_LEN=4000
|
||||
OUTPUT_LEN=16
|
||||
MIN_CACHE_HIT_PCT=0
|
||||
MAX_LATENCY_ALLOWED_MS=100000000000
|
||||
NUM_SEQS_LIST="128 256"
|
||||
NUM_BATCHED_TOKENS_LIST="512 1024 2048 4096"
|
||||
|
||||
LOG_FOLDER="$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
|
||||
mkdir -p $LOG_FOLDER
|
||||
mkdir -p $PROFILE_PATH
|
||||
|
||||
cd "$BASE/vllm"
|
||||
|
||||
@ -54,66 +60,36 @@ 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
|
||||
local max_num_seqs=$2
|
||||
local max_num_batched_tokens=$3
|
||||
local vllm_log=$4
|
||||
local profile_dir=$5
|
||||
|
||||
pkill -f vllm
|
||||
|
||||
pkill -if "vllm serve" || true
|
||||
|
||||
# Define the common arguments as a bash array.
|
||||
# Each argument and its value are separate elements.
|
||||
local common_args_array=(
|
||||
"$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_SERVER_DEV_MODE=1 VLLM_TORCH_PROFILER_DIR=$profile_dir \
|
||||
vllm serve "${common_args_array[@]}" > "$vllm_log" 2>&1 &
|
||||
else
|
||||
# Start server without profiling
|
||||
VLLM_SERVER_DEV_MODE=1 \
|
||||
vllm serve "${common_args_array[@]}" > "$vllm_log" 2>&1 &
|
||||
fi
|
||||
local server_pid=$!
|
||||
VLLM_USE_V1=1 VLLM_SERVER_DEV_MODE=1 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
|
||||
# This line checks whether the server is still alive or not,
|
||||
# since that we should always have permission to send signal to the server process.
|
||||
kill -0 $server_pid 2> /dev/null || break
|
||||
|
||||
for i in {1..60}; do
|
||||
RESPONSE=$(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
|
||||
@ -121,9 +97,8 @@ start_server() {
|
||||
sleep 10
|
||||
fi
|
||||
done
|
||||
|
||||
if (( ! server_started )); then
|
||||
echo "server did not start within 10 minutes or crashed. Please check server log at $vllm_log".
|
||||
echo "server did not start within 10 minutes. Please check server log at $vllm_log".
|
||||
return 1
|
||||
else
|
||||
return 0
|
||||
@ -139,11 +114,10 @@ run_benchmark() {
|
||||
echo "vllm_log: $vllm_log"
|
||||
echo
|
||||
rm -f $vllm_log
|
||||
pkill -if "vllm serve" || true
|
||||
pkill -f vllm
|
||||
|
||||
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 $vllm_log
|
||||
result=$?
|
||||
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: $max_num_batched_tokens"
|
||||
@ -151,19 +125,17 @@ run_benchmark() {
|
||||
echo "server started."
|
||||
fi
|
||||
echo
|
||||
|
||||
|
||||
echo "run benchmark test..."
|
||||
meet_latency_requirement=0
|
||||
# get a basic qps by using request-rate inf
|
||||
bm_log="$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 $OUTPUT_LEN \
|
||||
--ignore-eos \
|
||||
--disable-tqdm \
|
||||
@ -190,11 +162,11 @@ run_benchmark() {
|
||||
curl -X POST http://0.0.0.0:8004/reset_prefix_cache
|
||||
sleep 5
|
||||
bm_log="$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 $OUTPUT_LEN \
|
||||
--ignore-eos \
|
||||
--disable-tqdm \
|
||||
@ -223,7 +195,6 @@ run_benchmark() {
|
||||
best_max_num_seqs=$max_num_seqs
|
||||
best_num_batched_tokens=$max_num_batched_tokens
|
||||
best_goodput=$goodput
|
||||
best_request_rate=$request_rate
|
||||
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}"
|
||||
@ -232,9 +203,9 @@ 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 serve" || true
|
||||
pkill vllm
|
||||
sleep 10
|
||||
echo "===================="
|
||||
printf '=%.0s' $(seq 1 20)
|
||||
return 0
|
||||
}
|
||||
|
||||
@ -245,8 +216,7 @@ read -r -a num_batched_tokens_list <<< "$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 [[ "$result" -eq 0 ]]; then
|
||||
find_gpu_memory_utilization=1
|
||||
@ -269,45 +239,6 @@ for num_seqs in "${num_seqs_list[@]}"; do
|
||||
done
|
||||
done
|
||||
echo "finish permutations"
|
||||
echo "best_max_num_seqs: $best_max_num_seqs, best_num_batched_tokens: $best_num_batched_tokens, best_throughput: $best_throughput"
|
||||
echo "best_max_num_seqs: $best_max_num_seqs, best_num_batched_tokens: $best_num_batched_tokens, best_throughput: $best_throughput" >> "$RESULT"
|
||||
|
||||
# =================================================================================
|
||||
# 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 serve" || true
|
||||
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"
|
||||
@ -1,218 +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.
|
||||
|
||||
## Batched `auto_tune`
|
||||
|
||||
The `batch_auto_tune.sh` script allows you to run multiple `auto_tune.sh` experiments sequentially from a single configuration file. It iterates through a list of parameter sets, executes `auto_tune.sh` for each, and records the results back into the input file.
|
||||
|
||||
### Prerequisites
|
||||
|
||||
- **jq**: This script requires `jq` to parse the JSON configuration file.
|
||||
- **gcloud**: If you plan to upload results to Google Cloud Storage, the `gcloud` CLI must be installed and authenticated.
|
||||
|
||||
### How to Run
|
||||
|
||||
1. **Create a JSON configuration file**: Create a file (e.g., `runs_config.json`) containing an array of JSON objects. Each object defines the parameters for a single `auto_tune.sh` run.
|
||||
|
||||
2. **Execute the script**:
|
||||
|
||||
```bash
|
||||
bash batch_auto_tune.sh <path_to_json_file> [gcs_upload_path]
|
||||
```
|
||||
|
||||
- `<path_to_json_file>`: **Required.** Path to your JSON configuration file.
|
||||
- `[gcs_upload_path]`: **Optional.** A GCS path (e.g., `gs://my-bucket/benchmark-results`) where the detailed results and profiles for each run will be uploaded. If this is empty, the results will be available on the local filesystem (see the log for `RESULT_FILE=/path/to/results/file.txt`).
|
||||
|
||||
### Configuration File
|
||||
|
||||
The JSON configuration file should contain an array of objects. Each object's keys correspond to the configuration variables for `auto_tune.sh` (see the [Configuration table above](#configuration)). These keys will be converted to uppercase environment variables for each run.
|
||||
|
||||
Here is an example `runs_config.json` with two benchmark configurations:
|
||||
|
||||
```json
|
||||
[
|
||||
{
|
||||
"base": "/home/user",
|
||||
"model": "meta-llama/Llama-3.1-8B-Instruct",
|
||||
"system": "TPU", # OR GPU
|
||||
"tp": 8,
|
||||
"input_len": 128,
|
||||
"output_len": 2048,
|
||||
"max_model_len": 2300,
|
||||
"num_seqs_list": "128 256",
|
||||
"num_batched_tokens_list": "8192 16384"
|
||||
},
|
||||
{
|
||||
"base": "/home/user",
|
||||
"model": "meta-llama/Llama-3.1-70B-Instruct",
|
||||
"system": "TPU", # OR GPU
|
||||
"tp": 8,
|
||||
"input_len": 4000,
|
||||
"output_len": 16,
|
||||
"max_model_len": 4096,
|
||||
"num_seqs_list": "64 128",
|
||||
"num_batched_tokens_list": "4096 8192",
|
||||
"max_latency_allowed_ms": 500
|
||||
}
|
||||
]
|
||||
```
|
||||
|
||||
### Output
|
||||
|
||||
The script modifies the input JSON file in place, adding the results of each run to the corresponding object. The following fields are added:
|
||||
|
||||
- `run_id`: A unique identifier for the run, derived from the timestamp.
|
||||
- `status`: The outcome of the run (`SUCCESS`, `FAILURE`, or `WARNING_NO_RESULT_FILE`).
|
||||
- `results`: The content of the `result.txt` file from the `auto_tune.sh` run.
|
||||
- `gcs_results`: The GCS URL where the run's artifacts are stored (if a GCS path was provided).
|
||||
|
||||
A summary of successful and failed runs is also printed to the console upon completion.
|
||||
@ -1,128 +0,0 @@
|
||||
#!/bin/bash
|
||||
|
||||
INPUT_JSON="$1"
|
||||
GCS_PATH="$2" # Optional GCS path for uploading results for each run
|
||||
|
||||
SCRIPT_DIR=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)
|
||||
AUTOTUNE_SCRIPT="$SCRIPT_DIR/auto_tune.sh"
|
||||
|
||||
if [[ -z "$INPUT_JSON" ]]; then
|
||||
echo "Error: Input JSON file not provided."
|
||||
echo "Usage: $0 <path_to_json_file> [gcs_upload_path]"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
if [[ ! -f "$INPUT_JSON" ]]; then
|
||||
echo "Error: File not found at '$INPUT_JSON'"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
if ! command -v jq &> /dev/null; then
|
||||
echo "Error: 'jq' command not found. Please install jq to process the JSON input."
|
||||
exit 1
|
||||
fi
|
||||
|
||||
if [[ -n "$GCS_PATH" ]] && ! command -v gcloud &> /dev/null; then
|
||||
echo "Error: 'gcloud' command not found, but a GCS_PATH was provided."
|
||||
exit 1
|
||||
fi
|
||||
|
||||
SUCCESS_COUNT=0
|
||||
FAILURE_COUNT=0
|
||||
FAILED_RUNS=()
|
||||
SCRIPT_START_TIME=$(date +%s)
|
||||
|
||||
json_content=$(cat "$INPUT_JSON")
|
||||
if ! num_runs=$(echo "$json_content" | jq 'length'); then
|
||||
echo "Error: Invalid JSON in $INPUT_JSON. 'jq' failed to get array length." >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
echo "Found $num_runs benchmark configurations in $INPUT_JSON."
|
||||
echo "Starting benchmark runs..."
|
||||
echo "--------------------------------------------------"
|
||||
|
||||
for i in $(seq 0 $(($num_runs - 1))); do
|
||||
run_object=$(echo "$json_content" | jq ".[$i]")
|
||||
|
||||
RUN_START_TIME=$(date +%s)
|
||||
ENV_VARS_ARRAY=()
|
||||
# Dynamically create env vars from the JSON object's keys
|
||||
for key in $(echo "$run_object" | jq -r 'keys_unsorted[]'); do
|
||||
value=$(echo "$run_object" | jq -r ".$key")
|
||||
var_name=$(echo "$key" | tr '[:lower:]' '[:upper:]' | tr -cd 'A-Z0-9_')
|
||||
ENV_VARS_ARRAY+=("${var_name}=${value}")
|
||||
done
|
||||
|
||||
echo "Executing run #$((i+1))/$num_runs with parameters: ${ENV_VARS_ARRAY[*]}"
|
||||
|
||||
# Execute auto_tune.sh and capture output
|
||||
RUN_OUTPUT_FILE=$(mktemp)
|
||||
if env "${ENV_VARS_ARRAY[@]}" bash "$AUTOTUNE_SCRIPT" > >(tee -a "$RUN_OUTPUT_FILE") 2>&1; then
|
||||
STATUS="SUCCESS"
|
||||
((SUCCESS_COUNT++))
|
||||
else
|
||||
STATUS="FAILURE"
|
||||
((FAILURE_COUNT++))
|
||||
FAILED_RUNS+=("Run #$((i+1)): $(echo $run_object | jq -c .)")
|
||||
fi
|
||||
|
||||
RUN_OUTPUT=$(<"$RUN_OUTPUT_FILE")
|
||||
rm "$RUN_OUTPUT_FILE"
|
||||
|
||||
# Parse results and optionally upload them to GCS
|
||||
RUN_ID=""
|
||||
RESULTS=""
|
||||
GCS_RESULTS_URL=""
|
||||
if [[ "$STATUS" == "SUCCESS" ]]; then
|
||||
RESULT_FILE_PATH=$(echo "$RUN_OUTPUT" | grep 'RESULT_FILE=' | tail -n 1 | cut -d'=' -f2 | tr -s '/' || true)
|
||||
|
||||
if [[ -n "$RESULT_FILE_PATH" && -f "$RESULT_FILE_PATH" ]]; then
|
||||
RUN_ID=$(basename "$(dirname "$RESULT_FILE_PATH")")
|
||||
RESULT_DIR=$(dirname "$RESULT_FILE_PATH")
|
||||
RESULTS=$(cat "$RESULT_FILE_PATH")
|
||||
|
||||
if [[ -n "$GCS_PATH" ]]; then
|
||||
GCS_RESULTS_URL="${GCS_PATH}/${RUN_ID}"
|
||||
echo "Uploading results to GCS..."
|
||||
if gcloud storage rsync --recursive "$RESULT_DIR/" "$GCS_RESULTS_URL"; then
|
||||
echo "GCS upload successful."
|
||||
else
|
||||
echo "Warning: GCS upload failed for RUN_ID $RUN_ID."
|
||||
fi
|
||||
fi
|
||||
else
|
||||
echo "Warning: Could not find result file for a successful run."
|
||||
STATUS="WARNING_NO_RESULT_FILE"
|
||||
fi
|
||||
fi
|
||||
|
||||
# Add the results back into the JSON object for this run
|
||||
json_content=$(echo "$json_content" | jq --argjson i "$i" --arg run_id "$RUN_ID" --arg status "$STATUS" --arg results "$RESULTS" --arg gcs_results "$GCS_RESULTS_URL" \
|
||||
'.[$i] += {run_id: $run_id, status: $status, results: $results, gcs_results: $gcs_results}')
|
||||
|
||||
RUN_END_TIME=$(date +%s)
|
||||
echo "Run finished in $((RUN_END_TIME - RUN_START_TIME)) seconds. Status: $STATUS"
|
||||
echo "--------------------------------------------------"
|
||||
|
||||
# Save intermediate progress back to the file
|
||||
echo "$json_content" > "$INPUT_JSON.tmp" && mv "$INPUT_JSON.tmp" "$INPUT_JSON"
|
||||
|
||||
done
|
||||
|
||||
SCRIPT_END_TIME=$(date +%s)
|
||||
echo "All benchmark runs completed in $((SCRIPT_END_TIME - SCRIPT_START_TIME)) seconds."
|
||||
echo
|
||||
echo "====================== SUMMARY ======================"
|
||||
echo "Successful runs: $SUCCESS_COUNT"
|
||||
echo "Failed runs: $FAILURE_COUNT"
|
||||
echo "==================================================="
|
||||
|
||||
if [[ $FAILURE_COUNT -gt 0 ]]; then
|
||||
echo "Details of failed runs (see JSON file for full parameters):"
|
||||
for failed in "${FAILED_RUNS[@]}"; do
|
||||
echo " - $failed"
|
||||
done
|
||||
fi
|
||||
|
||||
echo "Updated results have been saved to '$INPUT_JSON'."
|
||||
@ -8,6 +8,7 @@ import sys
|
||||
import time
|
||||
import traceback
|
||||
from dataclasses import dataclass, field
|
||||
from typing import Optional, Union
|
||||
|
||||
import aiohttp
|
||||
import huggingface_hub.constants
|
||||
@ -27,13 +28,12 @@ class RequestFuncInput:
|
||||
prompt_len: int
|
||||
output_len: int
|
||||
model: str
|
||||
model_name: str | None = None
|
||||
logprobs: int | None = None
|
||||
extra_body: dict | None = None
|
||||
multi_modal_content: dict | list[dict] | None = None
|
||||
model_name: Optional[str] = None
|
||||
logprobs: Optional[int] = None
|
||||
extra_body: Optional[dict] = None
|
||||
multi_modal_content: Optional[dict] = None
|
||||
ignore_eos: bool = False
|
||||
language: str | None = None
|
||||
request_id: str | None = None
|
||||
language: Optional[str] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
@ -51,7 +51,7 @@ class RequestFuncOutput:
|
||||
|
||||
async def async_request_tgi(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
api_url = request_func_input.api_url
|
||||
assert api_url.endswith("generate_stream")
|
||||
@ -71,9 +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:
|
||||
@ -85,9 +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()
|
||||
@ -132,7 +127,7 @@ async def async_request_tgi(
|
||||
|
||||
async def async_request_trt_llm(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
api_url = request_func_input.api_url
|
||||
assert api_url.endswith("generate_stream")
|
||||
@ -150,9 +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
|
||||
|
||||
@ -160,9 +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()
|
||||
@ -203,7 +193,7 @@ async def async_request_trt_llm(
|
||||
|
||||
async def async_request_deepspeed_mii(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
api_url = request_func_input.api_url
|
||||
assert api_url.endswith(("completions", "profile")), (
|
||||
@ -221,8 +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
|
||||
@ -266,7 +254,7 @@ async def async_request_deepspeed_mii(
|
||||
|
||||
async def async_request_openai_completions(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
api_url = request_func_input.api_url
|
||||
assert api_url.endswith(("completions", "profile")), (
|
||||
@ -295,8 +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
|
||||
@ -366,7 +352,7 @@ async def async_request_openai_completions(
|
||||
|
||||
async def async_request_openai_chat_completions(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
api_url = request_func_input.api_url
|
||||
assert api_url.endswith(("chat/completions", "profile")), (
|
||||
@ -378,15 +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
|
||||
@ -409,8 +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
|
||||
@ -428,14 +404,8 @@ async def async_request_openai_chat_completions(
|
||||
chunk_bytes = chunk_bytes.strip()
|
||||
if not chunk_bytes:
|
||||
continue
|
||||
chunk_bytes = chunk_bytes.decode("utf-8")
|
||||
# NOTE: SSE comments (often used as pings) start with a colon.
|
||||
# These are not JSON data payload and should be skipped.
|
||||
if chunk_bytes.startswith(":"):
|
||||
continue
|
||||
|
||||
chunk = chunk_bytes.removeprefix("data: ")
|
||||
|
||||
chunk = chunk_bytes.decode("utf-8").removeprefix("data: ")
|
||||
if chunk != "[DONE]":
|
||||
timestamp = time.perf_counter()
|
||||
data = json.loads(chunk)
|
||||
@ -475,7 +445,7 @@ async def async_request_openai_chat_completions(
|
||||
|
||||
async def async_request_openai_audio(
|
||||
request_func_input: RequestFuncInput,
|
||||
pbar: tqdm | None = None,
|
||||
pbar: Optional[tqdm] = None,
|
||||
) -> RequestFuncOutput:
|
||||
# Lazy import without PlaceholderModule to avoid vllm dep.
|
||||
import soundfile
|
||||
@ -507,8 +477,6 @@ 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):
|
||||
@ -517,10 +485,7 @@ 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():
|
||||
@ -609,7 +574,7 @@ def get_tokenizer(
|
||||
tokenizer_mode: str = "auto",
|
||||
trust_remote_code: bool = False,
|
||||
**kwargs,
|
||||
) -> PreTrainedTokenizer | PreTrainedTokenizerFast:
|
||||
) -> Union[PreTrainedTokenizer, PreTrainedTokenizerFast]:
|
||||
if pretrained_model_name_or_path is not None and not os.path.exists(
|
||||
pretrained_model_name_or_path
|
||||
):
|
||||
|
||||
@ -1,74 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import gc
|
||||
|
||||
from benchmark_utils import TimeCollector
|
||||
from tabulate import tabulate
|
||||
|
||||
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
|
||||
1167
benchmarks/benchmark_dataset.py
Normal file
1167
benchmarks/benchmark_dataset.py
Normal file
File diff suppressed because it is too large
Load Diff
@ -1,17 +1,186 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import sys
|
||||
"""Benchmark the latency of processing a single batch of requests."""
|
||||
|
||||
import argparse
|
||||
import dataclasses
|
||||
import json
|
||||
import os
|
||||
import time
|
||||
from typing import Any, Optional
|
||||
|
||||
import numpy as np
|
||||
from tqdm import tqdm
|
||||
|
||||
import vllm.envs as envs
|
||||
from benchmark_utils import convert_to_pytorch_benchmark_format, write_to_json
|
||||
from vllm import LLM, SamplingParams
|
||||
from vllm.engine.arg_utils import EngineArgs
|
||||
from vllm.inputs import PromptType
|
||||
from vllm.sampling_params import BeamSearchParams
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
|
||||
|
||||
def save_to_pytorch_benchmark_format(
|
||||
args: argparse.Namespace, results: dict[str, Any]
|
||||
) -> None:
|
||||
pt_records = convert_to_pytorch_benchmark_format(
|
||||
args=args,
|
||||
metrics={"latency": results["latencies"]},
|
||||
extra_info={k: results[k] for k in ["avg_latency", "percentiles"]},
|
||||
)
|
||||
if pt_records:
|
||||
pt_file = f"{os.path.splitext(args.output_json)[0]}.pytorch.json"
|
||||
write_to_json(pt_file, pt_records)
|
||||
|
||||
|
||||
def main(args: argparse.Namespace):
|
||||
print(args)
|
||||
|
||||
engine_args = EngineArgs.from_cli_args(args)
|
||||
|
||||
# NOTE(woosuk): If the request cannot be processed in a single batch,
|
||||
# the engine will automatically process the request in multiple batches.
|
||||
llm = LLM(**dataclasses.asdict(engine_args))
|
||||
assert llm.llm_engine.model_config.max_model_len >= (
|
||||
args.input_len + args.output_len
|
||||
), (
|
||||
"Please ensure that max_model_len is greater than"
|
||||
" the sum of input_len and output_len."
|
||||
)
|
||||
|
||||
sampling_params = SamplingParams(
|
||||
n=args.n,
|
||||
temperature=1.0,
|
||||
top_p=1.0,
|
||||
ignore_eos=True,
|
||||
max_tokens=args.output_len,
|
||||
detokenize=not args.disable_detokenize,
|
||||
)
|
||||
print(sampling_params)
|
||||
dummy_prompt_token_ids = np.random.randint(
|
||||
10000, size=(args.batch_size, args.input_len)
|
||||
)
|
||||
dummy_prompts: list[PromptType] = [
|
||||
{"prompt_token_ids": batch} for batch in dummy_prompt_token_ids.tolist()
|
||||
]
|
||||
|
||||
def llm_generate():
|
||||
if not args.use_beam_search:
|
||||
llm.generate(dummy_prompts, sampling_params=sampling_params, use_tqdm=False)
|
||||
else:
|
||||
llm.beam_search(
|
||||
dummy_prompts,
|
||||
BeamSearchParams(
|
||||
beam_width=args.n,
|
||||
max_tokens=args.output_len,
|
||||
ignore_eos=True,
|
||||
),
|
||||
)
|
||||
|
||||
def run_to_completion(profile_dir: Optional[str] = None):
|
||||
if profile_dir:
|
||||
llm.start_profile()
|
||||
llm_generate()
|
||||
llm.stop_profile()
|
||||
else:
|
||||
start_time = time.perf_counter()
|
||||
llm_generate()
|
||||
end_time = time.perf_counter()
|
||||
latency = end_time - start_time
|
||||
return latency
|
||||
|
||||
print("Warming up...")
|
||||
for _ in tqdm(range(args.num_iters_warmup), desc="Warmup iterations"):
|
||||
run_to_completion(profile_dir=None)
|
||||
|
||||
if args.profile:
|
||||
profile_dir = envs.VLLM_TORCH_PROFILER_DIR
|
||||
print(f"Profiling (results will be saved to '{profile_dir}')...")
|
||||
run_to_completion(profile_dir=profile_dir)
|
||||
return
|
||||
|
||||
# Benchmark.
|
||||
latencies = []
|
||||
for _ in tqdm(range(args.num_iters), desc="Profiling iterations"):
|
||||
latencies.append(run_to_completion(profile_dir=None))
|
||||
latencies = np.array(latencies)
|
||||
percentages = [10, 25, 50, 75, 90, 99]
|
||||
percentiles = np.percentile(latencies, percentages)
|
||||
print(f"Avg latency: {np.mean(latencies)} seconds")
|
||||
for percentage, percentile in zip(percentages, percentiles):
|
||||
print(f"{percentage}% percentile latency: {percentile} seconds")
|
||||
|
||||
# Output JSON results if specified
|
||||
if args.output_json:
|
||||
results = {
|
||||
"avg_latency": np.mean(latencies),
|
||||
"latencies": latencies.tolist(),
|
||||
"percentiles": dict(zip(percentages, percentiles.tolist())),
|
||||
}
|
||||
with open(args.output_json, "w") as f:
|
||||
json.dump(results, f, indent=4)
|
||||
save_to_pytorch_benchmark_format(args, results)
|
||||
|
||||
|
||||
def create_argument_parser():
|
||||
parser = FlexibleArgumentParser(
|
||||
description="Benchmark the latency of processing a single batch of "
|
||||
"requests till completion."
|
||||
)
|
||||
parser.add_argument("--input-len", type=int, default=32)
|
||||
parser.add_argument("--output-len", type=int, default=128)
|
||||
parser.add_argument("--batch-size", type=int, default=8)
|
||||
parser.add_argument(
|
||||
"--n",
|
||||
type=int,
|
||||
default=1,
|
||||
help="Number of generated sequences per prompt.",
|
||||
)
|
||||
parser.add_argument("--use-beam-search", action="store_true")
|
||||
parser.add_argument(
|
||||
"--num-iters-warmup",
|
||||
type=int,
|
||||
default=10,
|
||||
help="Number of iterations to run for warmup.",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-iters", type=int, default=30, help="Number of iterations to run."
|
||||
)
|
||||
parser.add_argument(
|
||||
"--profile",
|
||||
action="store_true",
|
||||
help="profile the generation process of a single batch",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--output-json",
|
||||
type=str,
|
||||
default=None,
|
||||
help="Path to save the latency results in JSON format.",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--disable-detokenize",
|
||||
action="store_true",
|
||||
help=(
|
||||
"Do not detokenize responses (i.e. do not include "
|
||||
"detokenization time in the latency measurement)"
|
||||
),
|
||||
)
|
||||
|
||||
parser = EngineArgs.add_cli_args(parser)
|
||||
# V1 enables prefix caching by default which skews the latency
|
||||
# numbers. We need to disable prefix caching by default.
|
||||
parser.set_defaults(enable_prefix_caching=False)
|
||||
|
||||
return parser
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
print("""DEPRECATED: This script has been moved to the vLLM CLI.
|
||||
|
||||
Please use the following command instead:
|
||||
vllm bench latency
|
||||
|
||||
For help with the new command, run:
|
||||
vllm bench latency --help
|
||||
|
||||
Alternatively, you can run the new command directly with:
|
||||
python -m vllm.entrypoints.cli.main bench latency --help
|
||||
""")
|
||||
sys.exit(1)
|
||||
parser = create_argument_parser()
|
||||
args = parser.parse_args()
|
||||
if args.profile and not envs.VLLM_TORCH_PROFILER_DIR:
|
||||
raise OSError(
|
||||
"The environment variable 'VLLM_TORCH_PROFILER_DIR' is not set. "
|
||||
"Please set it to a valid path to use torch profiler."
|
||||
)
|
||||
main(args)
|
||||
|
||||
@ -1,213 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import gc
|
||||
import time
|
||||
from unittest import mock
|
||||
|
||||
import numpy as np
|
||||
from benchmark_utils import TimeCollector
|
||||
from tabulate import tabulate
|
||||
|
||||
from vllm.config import (
|
||||
CacheConfig,
|
||||
DeviceConfig,
|
||||
LoadConfig,
|
||||
ModelConfig,
|
||||
ParallelConfig,
|
||||
SchedulerConfig,
|
||||
SpeculativeConfig,
|
||||
VllmConfig,
|
||||
)
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
from vllm.v1.spec_decode.ngram_proposer import NgramProposer
|
||||
from vllm.v1.worker.gpu_input_batch import InputBatch
|
||||
from vllm.v1.worker.gpu_model_runner import GPUModelRunner
|
||||
|
||||
|
||||
def benchmark_propose(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 benchmark_batched_propose(args):
|
||||
NUM_SPECULATIVE_TOKENS_NGRAM = 10
|
||||
PROMPT_LOOKUP_MIN = 5
|
||||
PROMPT_LOOKUP_MAX = 15
|
||||
MAX_MODEL_LEN = int(1e7)
|
||||
DEVICE = current_platform.device_type
|
||||
|
||||
model_config = ModelConfig(model="facebook/opt-125m", runner="generate")
|
||||
|
||||
speculative_config = SpeculativeConfig(
|
||||
target_model_config=model_config,
|
||||
target_parallel_config=ParallelConfig(),
|
||||
method="ngram",
|
||||
num_speculative_tokens=NUM_SPECULATIVE_TOKENS_NGRAM,
|
||||
prompt_lookup_max=PROMPT_LOOKUP_MAX,
|
||||
prompt_lookup_min=PROMPT_LOOKUP_MIN,
|
||||
)
|
||||
|
||||
vllm_config = VllmConfig(
|
||||
model_config=model_config,
|
||||
cache_config=CacheConfig(),
|
||||
speculative_config=speculative_config,
|
||||
device_config=DeviceConfig(device=current_platform.device_type),
|
||||
parallel_config=ParallelConfig(),
|
||||
load_config=LoadConfig(),
|
||||
scheduler_config=SchedulerConfig(),
|
||||
)
|
||||
|
||||
# monkey patch vllm.v1.worker.gpu_model_runner.get_pp_group
|
||||
mock_pp_group = mock.MagicMock()
|
||||
mock_pp_group.world_size = 1
|
||||
with mock.patch(
|
||||
"vllm.v1.worker.gpu_model_runner.get_pp_group", return_value=mock_pp_group
|
||||
):
|
||||
runner = GPUModelRunner(vllm_config, DEVICE)
|
||||
|
||||
# hack max model len
|
||||
runner.max_model_len = MAX_MODEL_LEN
|
||||
runner.drafter.max_model_len = MAX_MODEL_LEN
|
||||
|
||||
dummy_input_batch = InputBatch(
|
||||
max_num_reqs=args.num_req,
|
||||
max_model_len=MAX_MODEL_LEN,
|
||||
max_num_batched_tokens=args.num_req * args.num_token,
|
||||
device=DEVICE,
|
||||
pin_memory=False,
|
||||
vocab_size=256000,
|
||||
block_sizes=[16],
|
||||
)
|
||||
dummy_input_batch._req_ids = list(str(id) for id in range(args.num_req))
|
||||
dummy_input_batch.spec_decode_unsupported_reqs = ()
|
||||
dummy_input_batch.num_tokens_no_spec = [args.num_token] * args.num_req
|
||||
dummy_input_batch.token_ids_cpu = np.random.randint(
|
||||
0, 20, (args.num_req, args.num_token)
|
||||
)
|
||||
|
||||
runner.input_batch = dummy_input_batch
|
||||
|
||||
sampled_token_ids = [[0]] * args.num_req
|
||||
|
||||
print("Starting benchmark")
|
||||
# first run is warmup so ignore it
|
||||
for _ in range(args.num_iteration):
|
||||
start = time.time()
|
||||
runner.drafter.propose(
|
||||
sampled_token_ids,
|
||||
dummy_input_batch.req_ids,
|
||||
dummy_input_batch.num_tokens_no_spec,
|
||||
dummy_input_batch.token_ids_cpu,
|
||||
dummy_input_batch.spec_decode_unsupported_reqs,
|
||||
)
|
||||
end = time.time()
|
||||
print(f"Iteration time (s): {end - start}")
|
||||
|
||||
|
||||
def invoke_main() -> None:
|
||||
parser = FlexibleArgumentParser(
|
||||
description="Benchmark the performance of N-gram speculative decode drafting"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--batched", action="store_true", help="consider time to prepare batch"
|
||||
)
|
||||
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()
|
||||
|
||||
if not args.batched:
|
||||
benchmark_propose(args)
|
||||
else:
|
||||
benchmark_batched_propose(args)
|
||||
|
||||
|
||||
"""
|
||||
# Example command lines:
|
||||
# time python3 benchmarks/benchmark_ngram_proposer.py
|
||||
# time python3 benchmarks/benchmark_ngram_proposer.py --batched --num-iteration 4 --num-token 1000000 --num-req 128
|
||||
""" # noqa: E501
|
||||
if __name__ == "__main__":
|
||||
invoke_main() # pragma: no cover
|
||||
@ -32,6 +32,7 @@ import dataclasses
|
||||
import json
|
||||
import random
|
||||
import time
|
||||
from typing import Optional
|
||||
|
||||
from transformers import PreTrainedTokenizerBase
|
||||
|
||||
@ -79,7 +80,7 @@ def sample_requests_from_dataset(
|
||||
num_requests: int,
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
input_length_range: tuple[int, int],
|
||||
fixed_output_len: int | None,
|
||||
fixed_output_len: Optional[int],
|
||||
) -> list[Request]:
|
||||
if fixed_output_len is not None and fixed_output_len < 4:
|
||||
raise ValueError("output_len too small")
|
||||
@ -127,7 +128,7 @@ def sample_requests_from_random(
|
||||
num_requests: int,
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
input_length_range: tuple[int, int],
|
||||
fixed_output_len: int | None,
|
||||
fixed_output_len: Optional[int],
|
||||
prefix_len: int,
|
||||
) -> list[Request]:
|
||||
requests = []
|
||||
|
||||
@ -7,6 +7,7 @@ import dataclasses
|
||||
import json
|
||||
import random
|
||||
import time
|
||||
from typing import Optional
|
||||
|
||||
from transformers import AutoTokenizer, PreTrainedTokenizerBase
|
||||
|
||||
@ -23,7 +24,7 @@ def sample_requests(
|
||||
dataset_path: str,
|
||||
num_requests: int,
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
fixed_output_len: int | None,
|
||||
fixed_output_len: Optional[int],
|
||||
) -> list[tuple[str, int, int, int]]:
|
||||
if fixed_output_len is not None and fixed_output_len < 4:
|
||||
raise ValueError("output_len too small")
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@ -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 \
|
||||
@ -31,19 +31,20 @@ import time
|
||||
import uuid
|
||||
import warnings
|
||||
from collections.abc import AsyncGenerator
|
||||
from contextlib import nullcontext
|
||||
from dataclasses import dataclass
|
||||
from typing import Optional
|
||||
|
||||
import datasets
|
||||
import numpy as np
|
||||
import pandas as pd
|
||||
from tqdm.asyncio import tqdm
|
||||
from transformers import PreTrainedTokenizerBase
|
||||
|
||||
from backend_request_func import (
|
||||
ASYNC_REQUEST_FUNCS,
|
||||
RequestFuncInput,
|
||||
RequestFuncOutput,
|
||||
)
|
||||
from tqdm.asyncio import tqdm
|
||||
from transformers import PreTrainedTokenizerBase
|
||||
|
||||
try:
|
||||
from vllm.transformers_utils.tokenizer import get_tokenizer
|
||||
@ -316,7 +317,7 @@ def calculate_metrics(
|
||||
tokenizer: PreTrainedTokenizerBase,
|
||||
selected_percentile_metrics: list[str],
|
||||
selected_percentiles: list[float],
|
||||
goodput_config_dict: dict[str, float] | None = None,
|
||||
goodput_config_dict: Optional[dict[str, float]] = None,
|
||||
) -> tuple[BenchmarkMetrics, list[int]]:
|
||||
actual_output_lens: list[int] = []
|
||||
total_input = 0
|
||||
@ -436,9 +437,9 @@ async def benchmark(
|
||||
selected_percentile_metrics: list[str],
|
||||
selected_percentiles: list[str],
|
||||
ignore_eos: bool,
|
||||
max_concurrency: int | None,
|
||||
max_concurrency: Optional[int],
|
||||
structured_output_ratio: float,
|
||||
goodput_config_dict: dict[str, float] | None = None,
|
||||
goodput_config_dict: Optional[dict[str, float]] = None,
|
||||
):
|
||||
if backend in ASYNC_REQUEST_FUNCS:
|
||||
request_func = ASYNC_REQUEST_FUNCS[backend]
|
||||
@ -448,8 +449,7 @@ async def benchmark(
|
||||
def prepare_extra_body(request) -> dict:
|
||||
extra_body = {}
|
||||
# Add the schema to the extra_body
|
||||
extra_body["structured_outputs"] = {}
|
||||
extra_body["structured_outputs"][request.structure_type] = request.schema
|
||||
extra_body[request.structure_type] = request.schema
|
||||
return extra_body
|
||||
|
||||
print("Starting initial single prompt test run...")
|
||||
@ -502,9 +502,15 @@ async def benchmark(
|
||||
|
||||
pbar = None if disable_tqdm else tqdm(total=len(input_requests))
|
||||
|
||||
semaphore = asyncio.Semaphore(max_concurrency) if max_concurrency else nullcontext()
|
||||
# This can be used once the minimum Python version is 3.10 or higher,
|
||||
# and it will simplify the code in limited_request_func.
|
||||
# semaphore = (asyncio.Semaphore(max_concurrency)
|
||||
# if max_concurrency else contextlib.nullcontext())
|
||||
semaphore = asyncio.Semaphore(max_concurrency) if max_concurrency else None
|
||||
|
||||
async def limited_request_func(request_func_input, pbar):
|
||||
if semaphore is None:
|
||||
return await request_func(request_func_input=request_func_input, pbar=pbar)
|
||||
async with semaphore:
|
||||
return await request_func(request_func_input=request_func_input, pbar=pbar)
|
||||
|
||||
@ -532,6 +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()
|
||||
|
||||
@ -549,10 +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))
|
||||
@ -650,20 +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
|
||||
|
||||
|
||||
@ -690,11 +692,11 @@ def evaluate(ret, args):
|
||||
return re.match(args.regex, actual) is not None
|
||||
|
||||
def _eval_correctness(expected, actual):
|
||||
if args.structure_type == "json":
|
||||
if args.structure_type == "guided_json":
|
||||
return _eval_correctness_json(expected, actual)
|
||||
elif args.structure_type == "regex":
|
||||
elif args.structure_type == "guided_regex":
|
||||
return _eval_correctness_regex(expected, actual)
|
||||
elif args.structure_type == "choice":
|
||||
elif args.structure_type == "guided_choice":
|
||||
return _eval_correctness_choice(expected, actual)
|
||||
else:
|
||||
return None
|
||||
@ -774,18 +776,18 @@ def main(args: argparse.Namespace):
|
||||
)
|
||||
|
||||
if args.dataset == "grammar":
|
||||
args.structure_type = "grammar"
|
||||
args.structure_type = "guided_grammar"
|
||||
elif args.dataset == "regex":
|
||||
args.structure_type = "regex"
|
||||
args.structure_type = "guided_regex"
|
||||
elif args.dataset == "choice":
|
||||
args.structure_type = "choice"
|
||||
args.structure_type = "guided_choice"
|
||||
else:
|
||||
args.structure_type = "json"
|
||||
args.structure_type = "guided_json"
|
||||
|
||||
if args.no_structured_output:
|
||||
args.structured_output_ratio = 0
|
||||
if args.save_results:
|
||||
result_file_name = f"{args.structured_output_ratio}so"
|
||||
result_file_name = f"{args.structured_output_ratio}guided"
|
||||
result_file_name += f"_{backend}"
|
||||
result_file_name += f"_{args.request_rate}qps"
|
||||
result_file_name += f"_{args.model.split('/')[-1]}"
|
||||
@ -903,13 +905,13 @@ def create_argument_parser():
|
||||
parser.add_argument(
|
||||
"--tokenizer",
|
||||
type=str,
|
||||
help="Name or path of the tokenizer, if not using the default tokenizer.",
|
||||
help="Name or path of the tokenizer, if not using the default tokenizer.", # noqa: E501
|
||||
)
|
||||
parser.add_argument(
|
||||
"--tokenizer-mode",
|
||||
type=str,
|
||||
default="auto",
|
||||
help="Name or path of the tokenizer, if not using the default tokenizer.",
|
||||
help="Name or path of the tokenizer, if not using the default tokenizer.", # noqa: E501
|
||||
)
|
||||
parser.add_argument(
|
||||
"--num-prompts",
|
||||
@ -992,7 +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".',
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user