Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 26, 2025

• Enable move warnings for torch targets (#149923)

  Show description for 79e8a69
  

  cyyever

  authored and

  pytorchmergebot

  committed
  79e8a69

  Copy full SHA for 79e8a69

• [MPS] Fix metal ops with different dtypes (#149974)

  Show description for de68ddc
  

  malfet

  authored and

  pytorchmergebot

  committed
  de68ddc

  Copy full SHA for de68ddc

• Skip cxxabi check for s390x (#149954)

  Show description for aa575ca
  

  2 peopleauthored and

  pytorchmergebot

  committed
  aa575ca

  Copy full SHA for aa575ca

• [ONNX] Clean up the diagnostics module (#149864)

  Show description for 6ae8eb8
  

  justinchuby

  authored and

  pytorchmergebot

  committed
  6ae8eb8

  Copy full SHA for 6ae8eb8

• Revert "[custom_ops][perf] Move expensive pytree traversals of tensors to C++ (#148555)"

  Show description for d256b2d
  pytorchmergebot

  committed
  d256b2d

  Copy full SHA for d256b2d

• Support None return type in torchbind and Add more AOTI torchbind e2e tests (#149749)

  Show description for 819b23e
  

  yushangdi

  authored and

  pytorchmergebot

  committed
  819b23e

  Copy full SHA for 819b23e

• [inductor] Fix division by zero error in fractional max (#148729)

  Show description for 71acb1b
  

  isuruf

  authored and

  pytorchmergebot

  committed
  71acb1b

  Copy full SHA for 71acb1b

• [CUDA]][SymmetricMemory] Interpret empty string as `std::nullopt` in `rendezvous` (#149793)

  Show description for 9108d15
  

  eqy

  authored and

  pytorchmergebot

  committed
  9108d15

  Copy full SHA for 9108d15

• Revert "[inductor] Fix mm logging for `torch._scaled_.mm` (#149967)"

  Show description for ab9ca6b
  pytorchmergebot

  committed
  ab9ca6b

  Copy full SHA for ab9ca6b

• [ROCm][TunableOp] TunableOp Context Manager for unit tests (#149930)

  Show description for 45b1173
  

  naromero77amd

  authored and

  pytorchmergebot

  committed
  45b1173

  Copy full SHA for 45b1173

• [inductor][triton 3.3] Fix cpp_wrapper w/ TMA in triton 3.3 (#149973)

  Show description for a8d0c5c
  

  davidberard98

  authored and

  pytorchmergebot

  committed
  a8d0c5c

  Copy full SHA for a8d0c5c

• Revert "cpp_wrapper: Fix even more tests (#147225)"

  Show description for 1b373f6
  pytorchmergebot

  committed
  1b373f6

  Copy full SHA for 1b373f6

• Revert "cpp_wrapper: precompile a few more commonly used headers, and improve RAIIPyObject interface (#149350)"

  Show description for 91bf925
  pytorchmergebot

  committed
  91bf925

  Copy full SHA for 91bf925

Commits on Mar 25, 2025

• Fix broken LazyLinear init (#149693)

  Show description for 3c85784
  

  Vincent Moens

  authored and

  pytorchmergebot

  committed
  3c85784

  Copy full SHA for 3c85784

• [inductor] Fix mm logging for `torch._scaled_.mm` (#149967)

  Show description for 661d74b
  

  YUNQIUGUO

  authored and

  pytorchmergebot

  committed
  661d74b

  Copy full SHA for 661d74b

• [ROCm] fix uninitialized warning in BFloat16.h (#149868)

  Show description for c05328e
  

  ethanwee1

  authored and

  pytorchmergebot

  committed
  c05328e

  Copy full SHA for c05328e

• [ROCm] missing AT_CUDA_CHECK for cub and SoftMax (#149883)

  Show description for 36eb64d
  

  ethanwee1

  authored and

  pytorchmergebot

  committed
  36eb64d

  Copy full SHA for 36eb64d

• [cuDNN][SDPA] cuDNN SDPA supports `head_dim <= 256` on `sm90` and `sm100` as of `9.5.1+` (#149904)

  Show description for de73790
  

  eqy

  authored and

  pytorchmergebot

  committed
  de73790

  Copy full SHA for de73790

• Fix #149806 : Fix path lookup in _preload_cuda_deps (#149808)

  Show description for 68b3273
  

  Divain

  authored and

  pytorchmergebot

  committed
  68b3273

  Copy full SHA for 68b3273

• [Submodule] [cpuinfo] cpuinfo update (#149305)

  Show description for ce54c43
  

  Ozan Aydin

  authored and

  pytorchmergebot

  committed
  ce54c43

  Copy full SHA for ce54c43

• [AOTInductor] Refine error message for dlopen in AOTInductor (#149812)

  Show description for feb503c
  

  muchulee8

  authored and

  pytorchmergebot

  committed
  feb503c

  Copy full SHA for feb503c

• [ROCm] build magma rocm and upload tarball (#149902)

  Show description for 0159f8e
  

  2 peopleauthored and

  pytorchmergebot

  committed
  0159f8e

  Copy full SHA for 0159f8e

• Revert "[ROCm] build magma rocm and upload tarball (#149902)"

  Show description for d3b7cf7
  pytorchmergebot

  committed
  d3b7cf7

  Copy full SHA for d3b7cf7

• [MPS/Inductor] Add support for chebyshev_polynomial_t. (#149928)

  Show description for e85ce64
  

  dcci

  authored and

  pytorchmergebot

  committed
  e85ce64

  Copy full SHA for e85ce64

• refresh results of benchmarks (#149936)

  Show description for 6c9d48b
  

  laithsakka

  authored and

  pytorchmergebot

  committed
  6c9d48b

  Copy full SHA for 6c9d48b

• Use statically known true in should_decompose_mm (#149950)

  Show description for 90110b0
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  90110b0

  Copy full SHA for 90110b0

• add some extra test oom skips for jetson due to lacking nvml support (#149587)

  Show description for ce3dc9e
  

  Fuzzkatt

  authored and

  pytorchmergebot

  committed
  ce3dc9e

  Copy full SHA for ce3dc9e

• test/test_cuda.py: rework TEST_PYNVML logic to make more sense, add not IS_JETSON condition (#149578)

  Show description for b562d22
  

  Fuzzkatt

  authored and

  pytorchmergebot

  committed
  b562d22

  Copy full SHA for b562d22

• [AOTInductor] Bug fix for freeing buffers when freeing multiple times (#149810)

  Show description for 12628ba
  

  muchulee8

  authored and

  pytorchmergebot

  committed
  12628ba

  Copy full SHA for 12628ba

• [ROCm] build magma rocm and upload tarball (#149902)

  Show description for bf8f4ef
  

  jeffdaily

  authored and

  pytorchmergebot

  committed
  bf8f4ef

  Copy full SHA for bf8f4ef

• [Bugfix] Add handling for buffer overrides (#149882)

  Show description for d1ff3ff
  

  Lucaskabela

  authored and

  pytorchmergebot

  committed
  d1ff3ff

  Copy full SHA for d1ff3ff

• Improve subproc autotuning implementation (#149700)

  Show description for 8cd6a13
  

  masnesral

  authored and

  pytorchmergebot

  committed
  8cd6a13

  Copy full SHA for 8cd6a13

• Revert "[ONNX] Clean up the diagnostics module (#149864)"

  Show description for 30e8be5
  pytorchmergebot

  committed
  30e8be5

  Copy full SHA for 30e8be5

• [dynamo] Fix handling of setattr with some tensor attributes (#149791)

  Show description for 1c98dc3
  

  StrongerXi

  authored and

  pytorchmergebot

  committed
  1c98dc3

  Copy full SHA for 1c98dc3

• cpp_wrapper: precompile a few more commonly used headers, and improve RAIIPyObject interface (#149350)

  Show description for 0de70fb
  

  benjaminglass1

  authored and

  pytorchmergebot

  committed
  0de70fb

  Copy full SHA for 0de70fb

Pagination

Previous

Next