Commits

Breadcrumbs

History for

onmain

User selector

All users

Datepicker

All time

Commit History

Commits on Apr 26, 2025

• Add option to use mempool on OOM (#151487)

  Show description for d22c4cc
  

  dsjohns2

  authored and

  pytorchmergebot

  committed
  d22c4cc

  Copy full SHA for d22c4cc

Commits on Mar 6, 2025

• Documents torch.cuda.MemPool API (#148374)

  Show description for 3960f97
  

  syed-ahmed

  authored and

  pytorchmergebot

  committed
  3960f97

  Copy full SHA for 3960f97

Commits on Feb 13, 2025

• Revert "Implement cuda graphs implementation of torch.cond and torch.while_loop (#140979)"

  Show description for 9a88300
  pytorchmergebot

  committed
  9a88300

  Copy full SHA for 9a88300

Commits on Feb 11, 2025

• Implement cuda graphs implementation of torch.cond and torch.while_loop (#140979)

  Show description for c7515da
  

  galv

  authored and

  pytorchmergebot

  committed
  c7515da

  Copy full SHA for c7515da

Commits on Feb 6, 2025

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for 9ee506b
  

  eqy

  authored and

  pytorchmergebot

  committed
  9ee506b

  Copy full SHA for 9ee506b

Commits on Jan 31, 2025

• Revert "[CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)"

  Show description for c3f71eb
  pytorchmergebot

  committed
  c3f71eb

  Copy full SHA for c3f71eb

Commits on Jan 30, 2025

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for e291724
  

  eqy

  authored and

  pytorchmergebot

  committed
  e291724

  Copy full SHA for e291724

Commits on Jan 27, 2025

• Revert "[CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)"

  Show description for c986eba
  pytorchmergebot

  committed
  c986eba

  Copy full SHA for c986eba

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for abf2898
  

  eqy

  authored and

  pytorchmergebot

  committed
  abf2898

  Copy full SHA for abf2898

Commits on Jan 23, 2025

• Revert "[CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)"

  Show description for dad9bc3
  pytorchmergebot

  committed
  dad9bc3

  Copy full SHA for dad9bc3

Commits on Jan 22, 2025

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for de945d7
  

  eqy

  authored and

  pytorchmergebot

  committed
  de945d7

  Copy full SHA for de945d7

Commits on Jan 16, 2025

• Revert "[CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)"

  Show description for 4ea1894
  pytorchmergebot

  committed
  4ea1894

  Copy full SHA for 4ea1894

Commits on Jan 15, 2025

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for a6763b7
  

  eqy

  authored and

  pytorchmergebot

  committed
  a6763b7

  Copy full SHA for a6763b7

Commits on Jan 14, 2025

• Revert "[CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)"

  Show description for 64bcf39
  pytorchmergebot

  committed
  64bcf39

  Copy full SHA for 64bcf39

Commits on Jan 11, 2025

• [CUDA][cuBLAS] Add fp16 accumulate option to cuBLAS/cuBLASLt (#144441)

  Show description for 388b75e
  

  eqy

  authored and

  pytorchmergebot

  committed
  388b75e

  Copy full SHA for 388b75e

Commits on Sep 17, 2024

• [PyTorch Pinned Allocator] Add support of background thread to process events (#135524)

  Show description for a575ce0
  

  banitag1

  authored and

  pytorchmergebot

  committed
  a575ce0

  Copy full SHA for a575ce0

• [PyTorch CUDA Allocator] Allow reuse of non-split blocks with better rounding (#136174)

  Show description for 48d18fb
  

  banitag1

  authored and

  pytorchmergebot

  committed
  48d18fb

  Copy full SHA for 48d18fb

Commits on Mar 27, 2024

• Graph-Safe RNG State Exchange for Tensor Parallelism (#114068)

  Show description for 249e65b
  

  eee4017

  authored and

  pytorchmergebot

  committed
  249e65b

  Copy full SHA for 249e65b

Commits on Mar 25, 2024

• Revert "Graph-Safe RNG State Exchange for Tensor Parallelism (#114068)"

  Show description for 4dc09d6
  pytorchmergebot

  committed
  4dc09d6

  Copy full SHA for 4dc09d6

Commits on Mar 21, 2024

• Graph-Safe RNG State Exchange for Tensor Parallelism (#114068)

  Show description for e9dcda5
  

  eee4017

  authored and

  pytorchmergebot

  committed
  e9dcda5

  Copy full SHA for e9dcda5

Commits on Feb 7, 2024

• Test seo torch cuda (#119324)

  Show description for 5ae6f6c
  

  svekars

  authored and

  pytorchmergebot

  committed
  5ae6f6c

  Copy full SHA for 5ae6f6c

Commits on Nov 2, 2023

• Add a note about performant record_stream use. (#112526)

  Show description for 09df6b7
  

  ezyang

  authored and

  pytorchmergebot

  committed
  09df6b7

  Copy full SHA for 09df6b7

Commits on Oct 19, 2023

• [DOCS][CUDA] Update TF32 docs for sm90 (#111337)

  Show description for 894b995
  

  eqy

  authored and

  pytorchmergebot

  committed
  894b995

  Copy full SHA for 894b995

Commits on Oct 6, 2023

• [CUDA Host Allocator] Add support of CudaHostRegister (#108488)

  Show description for 64583c4
  

  banitag1

  authored and

  pytorchmergebot

  committed
  64583c4

  Copy full SHA for 64583c4

Commits on Aug 24, 2023

• document memory snapshotting (#107660)

  Show description for 40cbda2
  

  zdevito

  authored and

  pytorchmergebot

  committed
  40cbda2

  Copy full SHA for 40cbda2

Commits on May 10, 2023

• [CUDA][CUBLAS] Fix BF16 reduced precision reduction note in docs (#101044)

  Show description for 33f3dca
  

  eqy

  authored and

  pytorchmergebot

  committed
  33f3dca

  Copy full SHA for 33f3dca

• [CUDA][CUBLAS] Add cuBLAS workspace allocation behavior to docs (#100919)

  Show description for 6e2efd1
  

  eqy

  authored and

  pytorchmergebot

  committed
  6e2efd1

  Copy full SHA for 6e2efd1

Commits on Jan 12, 2023

• Add `device_idx` to `free_fn` in `CUDAPluggableAllocator` (#91398)

  Show description for 07e595e
  

  Emilio Castillo

  authored and

  pytorchmergebot

  committed
  07e595e

  Copy full SHA for 07e595e

Commits on Dec 21, 2022

• [cuBLAS] Add an option to disable reduced precision reductions for BF16 GEMM (#89172)

  Show description for 8b617f8
  

  eqy

  authored and

  pytorchmergebot

  committed
  8b617f8

  Copy full SHA for 8b617f8

Commits on Nov 23, 2022

• Add Pluggable CUDA allocator backend (#86786)

  Show description for c9d4390
  

  Emilio Castillo

  authored and

  pytorchmergebot

  committed
  c9d4390

  Copy full SHA for c9d4390

Commits on Nov 4, 2022

• Modified roundup_power2_divisions to specify the number of divisions for each power of two interval (#87290)

  Show description for 5b767d4
  

  c-odrin

  authored and

  pytorchmergebot

  committed
  5b767d4

  Copy full SHA for 5b767d4

Commits on Oct 12, 2022

• Extend torch.cuda.is_available() to attempt an NVML-based CUDA availability assessment when explicitly requested by the user (#85951)

  Show description for ce56ee1
  

  speediedan

  authored and

  pytorchmergebot

  committed
  ce56ee1

  Copy full SHA for ce56ee1

• (Re-open) Adds cudaMallocAsync as an alternative backend for the CUDA allocator (#82682)

  Show description for 25725fd
  

  eqy

  authored and

  pytorchmergebot

  committed
  25725fd

  Copy full SHA for 25725fd

Commits on Oct 3, 2022

• Added roundup_bypass_threshold_mb knobs to the PyTorch Caching Allocator (#85940)

  Show description for d401732
  

  c-odrin

  authored and

  pytorchmergebot

  committed
  d401732

  Copy full SHA for d401732

Commits on Aug 26, 2022

• Fix small typo in cuda.rst (#84012)

  Show description for 089101f
  

  cjauvin

  authored and

  pytorchmergebot

  committed
  089101f

  Copy full SHA for 089101f

Pagination

Previous

Next