mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00

Files

Henry Tsang b878ca0c91 [cutlass backend] add fp8 to cutlass benchmark script (#155507 )

Summary:
Add fp8.

Right now FP8 only allows fast_accum.

Test Plan:
```
Experiment group: _scaled_mm (8192x8192, 8192x8192) torch.float8_e4m3fn
+-----------------------+--------------------+--------------------+----------------------+--------------------+
|         name          | forward_time (us)  | teraflops (TFLOPS) | compilation_time (s) | perf_over_aten (%) |
+-----------------------+--------------------+--------------------+----------------------+--------------------+
|         aten          | 967.1226739883423  | 1136.8895149998868 |  1.219131228979677   |         NA         |
|        triton         | 1764.6185159683228 |  623.08743664783   |  20.373826419003308  | 82.46067054670186  |
| triton_persistent_tma | 1769.0335512161255 | 621.5323768280928  |  20.48663099599071   | 82.91718297956578  |
|  cutlass_lvl_default  | 790.5075550079346  | 1390.8932568835019 |  13.788519630907103  | -18.26191482535096 |
|   cutlass_lvl_3332    | 803.7384748458862  | 1367.996757884245  |  226.81587297911756  | -16.89384434227684 |
+-----------------------+--------------------+--------------------+----------------------+--------------------+
```

Rollback Plan:

Differential Revision: D76310809

Pull Request resolved: https://github.com/pytorch/pytorch/pull/155507
Approved by: https://github.com/ColinPeppler

2025-06-13 05:11:15 +00:00

distributed/ddp

[BE] Remove outdated RPC benchmark (#146716 )

2025-03-29 04:44:36 +00:00

dynamo

refresh expected results (#155817 )

2025-06-12 19:14:21 +00:00

fastrnns

[BE]: Enable ruff rule SIM113 (#147290 )

2025-02-16 22:41:16 +00:00

framework_overhead_benchmark

Fix unused Python variables outside torch/ and test/ (#136359 )

2024-12-11 17:10:23 +00:00

functional_autograd_benchmark

Fix broken URLs (#152237 )

2025-04-27 09:56:42 +00:00

fuser

Fix unused Python variables outside torch/ and test/ (#136359 )

2024-12-11 17:10:23 +00:00

gpt_fast

[BE][CI] bump ruff to 0.9.2: multiline assert statements (#144546 )

2025-02-27 20:46:16 +00:00

inductor_backends

[cutlass backend] add fp8 to cutlass benchmark script (#155507 )

2025-06-13 05:11:15 +00:00

inference

[BE][Easy][3/19] enforce style for empty lines in import segments in benchmarks/ (#129754 )

2024-07-17 14:34:42 +00:00

instruction_counts

[BE][CI] bump ruff to 0.9.2: multiline assert statements (#144546 )

2025-02-27 20:46:16 +00:00

nested

Fix unused Python variables outside torch/ and test/ (#136359 )

2024-12-11 17:10:23 +00:00

operator_benchmark

add JSON output support for operator benchmark (#154410 )

2025-06-03 21:29:24 +00:00

overrides_benchmark

[BE][Easy][3/19] enforce style for empty lines in import segments in benchmarks/ (#129754 )

2024-07-17 14:34:42 +00:00

profiler_benchmark

Apply TorchFix TOR203 fixes (#143691 )

2024-12-23 18:21:03 +00:00

record_function_benchmark

[Caffe2]Remove Caffe2 scripts and benchmarks (#126747 )

2024-06-05 23:46:31 +00:00

serialization

Fix unused Python variables outside torch/ and test/ (#136359 )

2024-12-11 17:10:23 +00:00

sparse

Clean up conda usage in benchmark scripts (#152552 )

2025-04-30 21:27:29 +00:00

static_runtime

[3/N] Use internal linkage in C++ files (#151297 )

2025-05-05 17:48:39 +00:00

tensorexpr

Fix broken URLs (#152237 )

2025-04-27 09:56:42 +00:00

transformer

Add sparsity (#148513 )

2025-03-07 01:47:52 +00:00

compare-fastrnn-results.py

[BE][Easy][3/19] enforce style for empty lines in import segments in benchmarks/ (#129754 )

2024-07-17 14:34:42 +00:00

compare.sh

Benchmarks: add scripts for FastRNNs results comparison. (#44134 )

2020-09-03 13:44:42 -07:00

README.md

Removing conda references from PyTorch Docs (#152702 )

2025-05-20 20:33:28 +00:00

upload_scribe.py

Fix broken URLs (#152237 )

2025-04-27 09:56:42 +00:00

README.md

PyTorch Benchmarks

This folder contains scripts that produce reproducible timings of various PyTorch features.

It also provides mechanisms to compare PyTorch with other frameworks.

Setup environment

Make sure you're on a machine with CUDA, torchvision, and pytorch installed. Install in the following order:

# Install torchvision. It comes with the pytorch stable release binary
pip3 install torch torchvision

# Install the latest pytorch master from source.
# It should supersede the installation from the release binary.
cd $PYTORCH_HOME
python setup.py build develop

# Check the pytorch installation version
python -c "import torch; print(torch.__version__)"

Benchmark List

Please refer to each subfolder to discover each benchmark suite. Links are provided where descriptions exist: