main
Update dynamic_inductor_timm_training.csv (#167609)
add a curve for customized compilation in the kernel benchmarking scripts (#166697)
Enable all flake8-logging-format rules (#164655)
[DTensor] Add CPU instruction count benchmark for dispatch (#167394)
[BE][Easy] remove unused build-time dependency `astunparse` and change `astunparse.unparse` -> `ast.unparse` (#157907)
[doc] Rewrite benchmarks/dynamo/README.md (#115485)
Move TorchDynamo into PyTorch core (#86461)
[Benchmark] cleanup torchbench models (#164816)
Enable sample nightly PT2 benchmark on B200 (#158011)
[9/N] Apply ruff UP035 rule (#165515)
[ROCm][CI] dynamo benchmark repvgg_a2 is flaky (#167660)
[BE][CI] bump `ruff` to 0.9.0: string quote styles (#144569)
Improve benchmarks/dynamo:check_perf_csv output and failure summary (#161728)
[BE][Easy][3/19] enforce style for empty lines in import segments in `benchmarks/` (#129754)
[inductor] unittest for run2run determinism (#167482)
Enable PLC1802 on ruff (#165813)
[Benchmark] cleanup huggingface models (#164815)
Add mistral/gpt-oss to benchmarks (#163565)
[BE] Format `.ci/` / `.github/` / `benchmarks/` / `functorch/` / `tools/` / `torchgen/` with `ruff format` (#132577)
[torchbench][optimus] Add backend optimus (#167357)
[BugFix][Refactor] fix several instances which use f = open(...) without a corresponding f.close() (#167423)
Add benchmarks.py to run all benchmarks, add new file with all torchbench model names (#94146)
Utility for running delta comparisons between two flag configs (#95411)
Skip slow tests for aarch64-inductor-benchmarks (#158842)
[BE]: Update ruff to 0.11.8 (#153249)
[inductor] more aggressive mix order reduction (#166382)
[Benchmark] remove old timm models from benchmark (#164805)
[benchmark] Add more timm models (#165381)
[CD] Enable Inductor performance test for xpu (#166289)
Skip stable diffusion models in torchbench, get tests and benchmarks green (#167896)
[BE] fix ruff rule E226: add missing whitespace around operator in f-strings (#144415)