frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 12:54:11 +08:00
Code Issues Packages Projects Releases Wiki Activity
81,624 Commits 5,406 Branches 1,221 Tags
c17ba69ba5d3647b013ccd02be2a95c0ae3dc848
Commit Graph

625 Commits

Author SHA1 Message Date
atalman
c17ba69ba5 [submodule] Revert "Adds support for accelerated sorting with x86-simd-sort (#127936) (#141901) 
Looks like the original PR caused: https://github.com/pytorch/pytorch/issues/140590

Please see comment: https://github.com/pytorch/pytorch/issues/140590#issuecomment-2508704480

Pull Request resolved: https://github.com/pytorch/pytorch/pull/141901
Approved by: https://github.com/andrewor14, https://github.com/malfet
 2024-12-03 00:16:35 +00:00
cyy
0fca51bcc4 [11/N] Fix Wextra-semi warning (#140926) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140926
Approved by: https://github.com/ezyang
 2024-11-20 00:32:45 +00:00
Max Ren
cca34be584 Update XNNPACK Version (#139913) 
Updating XNNPACK Version to 4ea82e595b36106653175dcb04b2aa532660d0d8

submodule update
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139913
Approved by: https://github.com/digantdesai, https://github.com/huydhn
 2024-11-18 18:16:31 +00:00
Nathan Brown
a290c1d748 Fix building with system GLOO (#140275) 
Leverage existing FindGloo CMake module to locate system's library and headers. Add system's gloo headers to include path rather than the gloo from third party when USE_SYSTEM_GLOO is specified.

Fixes #140274

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140275
Approved by: https://github.com/malfet
 2024-11-11 22:58:39 +00:00
Yu, Guangye
8051ee802c Add XPU compiler version control in cmake to keep BC (#139258) 
# Motivation
This PR aims to maintain backward compatibility when building PyTorch XPU with the old and new compilers.

# Additional Context
The details are described here. The new compiler (2025.0.0) has some breaking changes compared with the old compiler(2024.1), for examples:
1. On Windows, sycl library is named `sycl7.lib` in the old compiler but is named `sycl.lib` in the new compiler.
2. On Linux, in order to support ABI=0, we have to link `libsycl-preview.so` in the old compiler but we could link `libsycl.so` in the new compiler to have the same ABI compatibility.
3. We added a macro `SYCL_COMPILER_VERSION` to support our new code has good backward compatibility with the old compiler. Now the new feature(Event elapsed_time, memory summary, and device architecture property) introduced by the new compiler will be controlled within the macro `SYCL_COMPILER_VERSION`.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139258
Approved by: https://github.com/EikanWang, https://github.com/atalman, https://github.com/gujinghui
 2024-11-09 13:31:21 +00:00
Matthew Sterrett
7e65060410 Adds support for accelerated sorting with x86-simd-sort (#127936) 
Adds x86-simd-sort as a submodule to accelerate sorting for 32-bit and 64-bit datatypes when AVX2 or AVX512 are available.

For contiguous data, this can be over a 10x speedup for large arrays. For discontiguous data, it can give over a 4x speedup with larger arrays. These benchmarks were gathered on a Skylake system (7900x), limited to 8 threads.

<details>
<summary><b>Contiguous Benchmarks</b></summary>

```
float32, normally distributed (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             7.150844336    6.886271477    7.132277489    1.038420335    1.002603214
128            9.208030939    8.478154898    7.846915245    1.086089019    1.173458697
1024           37.79037627    23.60707456    16.44122627    1.600807257    2.298513241
10000          714.7355628    203.9921844    105.5683001    3.503739934    6.770361577
100000         8383.074408    721.6333354    465.3709247    11.61680593    18.01374766
1000000        97124.31945    5632.054572    3920.148401    17.24491803    24.77567416
10000000       1161974.907    86070.48988    71533.82301    13.50027063    16.24371323

int32_t, uniformly distributed (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             7.203208685    6.92212224     7.014458179    1.040606975    1.026908779
128            8.972388983    8.195516348    7.592543125    1.094792396    1.18173698
1024           32.77489477    23.6874548     15.36617105    1.383639359    2.132925285
10000          607.8824128    193.3402024    99.25090471    3.144107667    6.124703997
100000         523.9384684    608.1836536    442.3166784    0.861480682    1.184532472
1000000        5211.348627    5271.598405    3518.861883    0.988570871    1.480975611
10000000       133853.6263    81463.05084    67852.97394    1.643120714    1.972700952
```

</details>

Note that the int32_t sort is accelerated by FBGEMM's radix sort for larger arrays, but this only handles contiguous data and in one sorting direction.

<details>
<summary><b>Discontiguous Benchmarks</b></summary>

```
float, normal distributed, discontiguous in sorted dimension (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             3.836543679    4.011214256    3.84376061     0.956454439    0.99812243
128            5.755310194    5.755723127    4.820394962    0.999928257    1.193949923
1024           49.46946019    24.78790785    15.47874362    1.995709379    3.195960952
10000          665.2505291    236.6165959    143.9490662    2.811512551    4.621429974
100000         4328.002203    1329.001212    818.3516414    3.256582586    5.288682743
1000000        47651.5018     16693.72045    11827.39551    2.854456677    4.028909133
10000000       556655.1288    236252.6258    184215.9828    2.356185998    3.021752621

int32_t, uniformly distributed, discontiguous in sorted dimension  (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             3.817994356    3.878117442    3.770039797    0.984496837    1.012719908
128            5.578731397    5.577152082    4.716770534    1.000283176    1.182743862
1024           43.3412619     23.61275801    14.55446819    1.835501887    2.977866408
10000          634.3997478    224.4322851    133.9518324    2.826686667    4.736028889
100000         4084.358152    1292.363303    781.7867576    3.16037924     5.22438902
1000000        46262.20465    16608.35284    11367.51817    2.785478192    4.06968381
10000000       541231.9104    235185.1861    180249.9294    2.301301028    3.002674742
```

</details>

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127936
Approved by: https://github.com/jgong5, https://github.com/peterbell10, https://github.com/sanchitintel
 2024-11-02 02:14:01 +00:00
Irem Yuksel
b021486405 Enable Windows Arm64 (#133088) 
This PR enables Pytorch for Windows on Arm64 - CPU only.
Currently, there aren't any checks in place to build and test for Windows on Arm64, but we're working to implement those as soon as possible.
We recommend using [Arm Performance Libraries (APL)](https://developer.arm.com/Tools%20and%20Software/Arm%20Performance%20Libraries) as a BLAS option, which is introduced in this PR.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/133088
Approved by: https://github.com/malfet

Co-authored-by: cristian panaite <panaite.cristian2000@gmail.com>
Co-authored-by: Stefan-Alin Pahontu <56953855+alinpahontu2912@users.noreply.github.com>
Co-authored-by: Ozan Aydin <148207261+ozanMSFT@users.noreply.github.com>
 2024-10-24 16:10:44 +00:00
cyy
af8bd323e8 Remove legacy Caffe2 pthreadpool from CMake (#134936) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/134936
Approved by: https://github.com/ezyang
 2024-10-17 05:22:08 +00:00
Nichols A. Romero
e8f1dd6ba0 Fix hardcoded ROCm paths in Caffe2Targets.cmake (#136283) 
Fixes #131701

Use CMake imported targets more consistently to eliminate hardcode paths.

Here is the new relevant sections of Caffe2Targets.cmake:
```
set_target_properties(c10_hip PROPERTIES
  INTERFACE_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include"
  INTERFACE_LINK_LIBRARIES "c10;hip::amdhip64"
)
```

```
set_target_properties(torch_hip PROPERTIES
  INTERFACE_COMPILE_DEFINITIONS "USE_C10D_NCCL"
  INTERFACE_COMPILE_OPTIONS "-fPIC;-D__HIP_PLATFORM_AMD__=1;-DCUDA_HAS_FP16=1;-DUSE_ROCM;-D__HIP_NO_HALF_OPERATORS__=1;-D__HIP_NO_HALF_CONVERSIONS__=1;-DTORCH_HIP_VERSION=602;-Wno-shift-count-negative;-Wno-shift-count-overflow;-Wno-duplicate-decl-specifier;-DCAFFE2_USE_MIOPEN;-DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_HIP;-std=c++17;-DHIPBLAS_V2;-DHIP_NEW_TYPE_ENUMS"
  INTERFACE_INCLUDE_DIRECTORIES "${_IMPORT_PREFIX}/include"
  INTERFACE_LINK_LIBRARIES "c10_hip;torch_cpu_library;hip::amdhip64;MIOpen;hiprtc::hiprtc;roc::hipblaslt;roc::hipblas;hip::hipfft;hip::hiprand;roc::hipsparse;roc::hipsolver"
)
```

HIPCUB dependency was not actually used; which is why it is removed here as the imported target had undesirable side effects.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/136283
Approved by: https://github.com/jeffdaily, https://github.com/Skylion007, https://github.com/jithunnair-amd, https://github.com/atalman
 2024-09-26 00:34:43 +00:00
PyTorch MergeBot
0e19522122 Revert "Adds support for accelerated sorting with x86-simd-sort (#127936)" 
This reverts commit 239a9ad65eebf93dcf9bb108a5129d4160b12c86.

Reverted https://github.com/pytorch/pytorch/pull/127936 on behalf of https://github.com/atalman due to test/test_sort_and_select.py::TestSortAndSelectCPU::test_sort_discontiguous_slow_cpu_float32 [GH job link](https://github.com/pytorch/pytorch/actions/runs/10994904767/job/30525578456) [HUD commit link](239a9ad65e) ([comment](https://github.com/pytorch/pytorch/pull/127936#issuecomment-2368522316))
 2024-09-23 14:52:23 +00:00
cyy
c459430558 Pass Werror to CUDA host compiler (#130213) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/130213
Approved by: https://github.com/ezyang
 2024-09-21 08:01:06 +00:00
Matthew Sterrett
239a9ad65e Adds support for accelerated sorting with x86-simd-sort (#127936) 
Adds x86-simd-sort as a submodule to accelerate sorting for 32-bit and 64-bit datatypes when AVX2 or AVX512 are available.

For contiguous data, this can be over a 10x speedup for large arrays. For discontiguous data, it can give over a 4x speedup with larger arrays. These benchmarks were gathered on a Skylake system (7900x), limited to 8 threads.

<details>
<summary><b>Contiguous Benchmarks</b></summary>

```
float32, normally distributed (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             7.150844336    6.886271477    7.132277489    1.038420335    1.002603214
128            9.208030939    8.478154898    7.846915245    1.086089019    1.173458697
1024           37.79037627    23.60707456    16.44122627    1.600807257    2.298513241
10000          714.7355628    203.9921844    105.5683001    3.503739934    6.770361577
100000         8383.074408    721.6333354    465.3709247    11.61680593    18.01374766
1000000        97124.31945    5632.054572    3920.148401    17.24491803    24.77567416
10000000       1161974.907    86070.48988    71533.82301    13.50027063    16.24371323

int32_t, uniformly distributed (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             7.203208685    6.92212224     7.014458179    1.040606975    1.026908779
128            8.972388983    8.195516348    7.592543125    1.094792396    1.18173698
1024           32.77489477    23.6874548     15.36617105    1.383639359    2.132925285
10000          607.8824128    193.3402024    99.25090471    3.144107667    6.124703997
100000         523.9384684    608.1836536    442.3166784    0.861480682    1.184532472
1000000        5211.348627    5271.598405    3518.861883    0.988570871    1.480975611
10000000       133853.6263    81463.05084    67852.97394    1.643120714    1.972700952
```

</details>

Note that the int32_t sort is accelerated by FBGEMM's radix sort for larger arrays, but this only handles contiguous data and in one sorting direction.

<details>
<summary><b>Discontiguous Benchmarks</b></summary>

```
float, normal distributed, discontiguous in sorted dimension (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             3.836543679    4.011214256    3.84376061     0.956454439    0.99812243
128            5.755310194    5.755723127    4.820394962    0.999928257    1.193949923
1024           49.46946019    24.78790785    15.47874362    1.995709379    3.195960952
10000          665.2505291    236.6165959    143.9490662    2.811512551    4.621429974
100000         4328.002203    1329.001212    818.3516414    3.256582586    5.288682743
1000000        47651.5018     16693.72045    11827.39551    2.854456677    4.028909133
10000000       556655.1288    236252.6258    184215.9828    2.356185998    3.021752621

int32_t, uniformly distributed, discontiguous in sorted dimension  (in microseconds)
size           Default        AVX2           AVX512         Default/AVX2   Default/AVX512
16             3.817994356    3.878117442    3.770039797    0.984496837    1.012719908
128            5.578731397    5.577152082    4.716770534    1.000283176    1.182743862
1024           43.3412619     23.61275801    14.55446819    1.835501887    2.977866408
10000          634.3997478    224.4322851    133.9518324    2.826686667    4.736028889
100000         4084.358152    1292.363303    781.7867576    3.16037924     5.22438902
1000000        46262.20465    16608.35284    11367.51817    2.785478192    4.06968381
10000000       541231.9104    235185.1861    180249.9294    2.301301028    3.002674742
```

</details>

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127936
Approved by: https://github.com/jgong5, https://github.com/peterbell10
 2024-09-20 21:19:33 +00:00
min-jean-cho
416a7894fe [Windows][XPU] Disable Kineto PTI on Windows only (#134620) 
Disable Kineto + XPU PTI on Windows only.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134620
Approved by: https://github.com/guangyey, https://github.com/malfet

Co-authored-by: Nikita Shulga <2453524+malfet@users.noreply.github.com>
 2024-08-29 20:58:55 +00:00
Gregory Comer
3b40b07efb Update PyTorch for XNNPACK 87ee0b4 (#134518) 
Summary: Update XNNPACK library version.

Test Plan: Combined diff CI is clean: D61586079 (all changes, has to be split out for export).

Differential Revision: D61822610

Pull Request resolved: https://github.com/pytorch/pytorch/pull/134518
Approved by: https://github.com/mcr229
 2024-08-28 19:24:04 +00:00
Xinya Zhang
5fd670e0ef [ROCM] Properly disable Flash Attention/Efficient Attention with environment variables (#133866) 
Now `USE_FLASH_ATTENTION=0 USE_MEM_EFF_ATTENTION=0 python setup.py` can compile correctly

Fixes #125230

Pull Request resolved: https://github.com/pytorch/pytorch/pull/133866
Approved by: https://github.com/jithunnair-amd, https://github.com/jeffdaily, https://github.com/malfet
 2024-08-27 18:24:29 +00:00
Mikayla Gawarecki
018e48c337 [Reland] Add wrappers for synchronous GPUDirect Storage APIs (#133489) 
Reland #130633

USE_CUFILE turned off by default in this version
Pull Request resolved: https://github.com/pytorch/pytorch/pull/133489
Approved by: https://github.com/albanD
 2024-08-15 17:11:52 +00:00
PyTorch MergeBot
fa1d7b0262 Revert "Remove unused Caffe2 macros (#132979)" 
This reverts commit da65cfbdea4f1f2176f6242004bda940a24f9ddb.

Reverted https://github.com/pytorch/pytorch/pull/132979 on behalf of https://github.com/ezyang due to these are apparently load bearing internally ([comment](https://github.com/pytorch/pytorch/pull/132979#issuecomment-2284666332))
 2024-08-12 18:34:56 +00:00
cyy
da65cfbdea Remove unused Caffe2 macros (#132979) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/132979
Approved by: https://github.com/ezyang
 2024-08-09 04:48:20 +00:00
cyy
05e8e87a69 [Submodule] Remove foxi (#132976) 
It is not used after removal of Caffe2 code.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132976
Approved by: https://github.com/ezyang
 2024-08-09 03:46:52 +00:00
Chen, Zejun
26b0011fb8 [XPU][Kineto Submodule] Introduce kineto-based XPU profiler (#130811) 
As XPU became a PyTorch built-in device, the profiler support is indispensable part of functionality completeness. This PR is associated with the PR to introduce XPU profiler plugin into the kineto. When USE_XPU is enabled, the LIBKINETO_NOXPUPTI option will be suppressed accordingly, which allows kineto to build with XPU profiler plugin.

Associated PR to introduce kineto-based XPU profiler into kineto:
https://github.com/pytorch/kineto/pull/961

Also updates the Kineto Submodule to include XPU changes.

Co-authored-by: Aaron Enye Shi <enye.shi@gmail.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/130811
Approved by: https://github.com/aaronenyeshi
 2024-08-07 18:41:37 +00:00
Yu, Guangye
92bebb46fa Support XPU ABI=0 build (#130110) 
# Motivation
This PR intends to support ABI=0 build for XPU backend.

# Additional Context
The major change is adding a compilation option `-D__INTEL_PREVIEW_BREAKING_CHANGES` for the host compiler(gcc) and `-fpreview-breaking-changes` for XPU device kernel code compiler(icpx), why?
Because we use
- gcc to compile host code and link SYCL runtime. So we need to pass `-D__INTEL_PREVIEW_BREAKING_CHANGES` to tell the host compiler invoking the ABI-neutral API included in SYCL. And
- use icpx to compile device kernel code and link SYCL runtime. So we need to pass `-fpreview-breaking-changes` to tell the device kernel compiler building ABI-neutral code. Besides,
- `libsycl-preview.so` is an ABI-neutral library but `libsycl.so` is not.

This PR depends on https://github.com/pytorch/pytorch/pull/131643.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/130110
Approved by: https://github.com/EikanWang, https://github.com/gujinghui, https://github.com/albanD
 2024-08-01 21:42:14 +00:00
PyTorch MergeBot
e191b83462 Revert "Add wrappers for synchronous GPUDirect Storage APIs (#130633)" 
This reverts commit 709ddf7a9dcfa1268848b72f6f56b55afa6728d6.

Reverted https://github.com/pytorch/pytorch/pull/130633 on behalf of https://github.com/clee2000 due to still failing internally D60265673 ([comment](https://github.com/pytorch/pytorch/pull/130633#issuecomment-2253239607))
 2024-07-26 18:08:20 +00:00
Mikayla Gawarecki
709ddf7a9d Add wrappers for synchronous GPUDirect Storage APIs (#130633) 
Based in part on https://github.com/NVIDIA/apex/pull/1774

Differential Revision: [D60155434](https://our.internmc.facebook.com/intern/diff/D60155434)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/130633
Approved by: https://github.com/albanD
 2024-07-25 22:23:38 +00:00
PyTorch MergeBot
e4b5645f83 Revert "Add wrappers for synchronous GPUDirect Storage APIs (#130633)" 
This reverts commit 5b5e0698a5f560decb9bbdd150ed7b0622eb7777.

Reverted https://github.com/pytorch/pytorch/pull/130633 on behalf of https://github.com/clee2000 due to breaking a lot of jobs and build rules internally D60085885, possibly needs to update some bazel build? ([comment](https://github.com/pytorch/pytorch/pull/130633#issuecomment-2245806738))
 2024-07-23 17:19:34 +00:00
Mikayla Gawarecki
5b5e0698a5 Add wrappers for synchronous GPUDirect Storage APIs (#130633) 
Based in part on https://github.com/NVIDIA/apex/pull/1774

Pull Request resolved: https://github.com/pytorch/pytorch/pull/130633
Approved by: https://github.com/albanD
 2024-07-22 14:51:24 +00:00
cyy
a6345d3477 [CMake] [3/N] Remove unused code (#130322) 
Some functions used by Caffe2 were removed along with some outdated checks. Follows #130006.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/130322
Approved by: https://github.com/r-barnes
 2024-07-09 19:33:33 +00:00
Yichen Yan
953c6476bd [CMAKE] Look for Development.Module instead of Development (#129669) 
Based on the [cmake issue](https://gitlab.kitware.com/cmake/cmake/-/issues/23716) and [manylinux issue](https://github.com/pypa/manylinux/issues/1347), when building a python module, it should find the `Development.Module` module, not `Development`, which includes `Development.Module` and `Development.Embed`, and will expect the shared python library only. After this PR and before #124613, pytorch could be built with a static libpython (e.g. in manylinux).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/129669
Approved by: https://github.com/malfet
 2024-07-09 09:16:43 +00:00
Shivam Raikundalia
a21d4363d2 [Profiler] Remove all instances of TMP_USE_TSC_AS_TIMESTAMP (#129973) 
Summary: Now that D56584521 is in, we can remove all insteances of TMP_USE_TSC_AS_TIMESTAMP

Test Plan:
Ran resnet. Trace looks good
https://www.internalfb.com/intern/perfdoctor/trace_view?filepath=tree/traces/dynocli/devvm2185.cco0.facebook.com/rank-0.Jun_27_14_46_01.1967733.pt.trace.json.gz&bucket=gpu_traces

Reviewed By: aaronenyeshi, swolchok

Differential Revision: D59132793

Pull Request resolved: https://github.com/pytorch/pytorch/pull/129973
Approved by: https://github.com/aaronenyeshi
 2024-07-03 19:28:52 +00:00
cyy
46366888d7 Remove outdated CMake code (#129851) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/129851
Approved by: https://github.com/ezyang
 2024-07-02 00:40:37 +00:00
Shivam Raikundalia
1d0efedc85 [Profiler] Add TSC Clock Callback to CUPTI (#125036) 
Summary:
Right now we use the default clock for CUPTI which is not monotonic nor particularly fast. We have already added the Kineto side of the implementation here: https://www.internalfb.com/diff/D56525885

This diff only adds the compile flags such that the TSC format is used and sets the converter using a libkineto call in the profiler

Test Plan:
Obtained following trace using resnet test:
https://www.internalfb.com/intern/perfdoctor/trace_view?filepath=tree/traces/dynocli/devvm2185.cco0.facebook.com/rank-0.Apr_25_11_03_18.3862943.pt.trace.json.gz&bucket=gpu_traces

TBD: Add benchmarks

Differential Revision: D56584521

Pull Request resolved: https://github.com/pytorch/pytorch/pull/125036
Approved by: https://github.com/aaronenyeshi
 2024-06-27 21:07:43 +00:00
Chirag Pandya
64f1111d38 Expose nholmann json to torch (#129570) 
Summary:

Expose nlohmann json library so that it can be used from inside Pytorch. The library already exists in the `third_party` directory. This PR is making `nlohmann/json.hpp` header available to be used from `torch.distributed`.
The next PR makes actual use of this header.

imported-using-ghimport

Test Plan: Imported from OSS

Reviewed By: malfet

Differential Revision: D59035246

Pulled By: c-p-i-o

Pull Request resolved: https://github.com/pytorch/pytorch/pull/129570
Approved by: https://github.com/d4l3k, https://github.com/malfet
 2024-06-26 21:59:26 +00:00
cyy
479ce5e2f4 Remove outdated CUDA code from CMake (#128801) 
It's possible to simplify some CUDA handling logic in CMake.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/128801
Approved by: https://github.com/r-barnes, https://github.com/malfet
 2024-06-21 15:00:00 +00:00
sdp
b4a0161449 Build SYCL kernels for ATen XPU ops on Native Windows (take 2) (#127390) 
Original PR https://github.com/pytorch/pytorch/pull/126725 is closed due to bad rebase.

-------
As proposed in https://github.com/pytorch/pytorch/issues/126719, we are enabling PyTorch XPU on Native Windows on Intel GPU.

This PR  enables XPU build on Windows as the first step of #126719:

- Enable `USE_XPU` build on Windows using MSVC as host compiler. The use of MSVC as host compiler seamlessly aligns with the existing PyTorch build on Windows.
- Build oneDNN GPU library on Windows.

Co-authored-by: Yu, Guangye <guangye.yu@intel.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127390
Approved by: https://github.com/guangyey, https://github.com/EikanWang, https://github.com/gujinghui, https://github.com/ezyang
 2024-06-06 01:41:06 +00:00
cyy
df75a9dc80 Remove Caffe2/onnx (#127991) 
Remove Caffe2/onnx since it is not used. Other tiny fixes are also applied.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127991
Approved by: https://github.com/ezyang
 2024-06-05 15:10:12 +00:00
Tristan Rice
597922ba21 Reapply "distributed debug handlers (#126601)" (#127805) 
This reverts commit 7646825c3eb687030c4f873b01312be0eed80174.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127805
Approved by: https://github.com/PaliC
 2024-06-04 19:44:30 +00:00
atalman
53f001c599 Revert "correct BLAS input (#126200)" (#127762) 
This reverts commit ea13e9a097aaa875a2b404822579b7f8b62ea291.

Looks like this could have caused: https://github.com/pytorch/pytorch/actions/runs/9346105069/job/25722431775#step:17:984

Aarch64 tests failures:
```
+ echo 'Checking that MKLDNN is available on aarch64'
Checking that MKLDNN is available on aarch64
+ pushd /tmp
/tmp /
+ python -c 'import torch; exit(0 if torch.backends.mkldnn.is_available() else 1)'
Error: Process completed with exit code 1.
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127762
Approved by: https://github.com/PaliC, https://github.com/malfet
 2024-06-03 15:49:48 +00:00
cyy
4e7f497bb3 [Submodule] Remove ios-cmake (#127694) 
It has not been updated for a long time and CI iOS builds don't rely on it.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127694
Approved by: https://github.com/ezyang
 2024-06-02 04:40:21 +00:00
Robert Mast
ea13e9a097 correct BLAS input (#126200) 
Fixes #32407

With this little correction to Dependencies.cmake it is possible to build an MKL-free version of Pytorch up from version v2.0.0 by explicitly choosing another MKL-free BLAS.

This pullrequest fulfills the "if not already present" part of the original comment in  Dependencies.cmake:
"setting default preferred BLAS options if not already present."

It's tested with this Action-.yml:
```
name: Build PyTorch v2.0.0 without AVX

on:
  push:
    branches:
      - v2.0.0
  pull_request:
    branches:
      - v2.0.0

jobs:
  build:
    runs-on: ubuntu-20.04
    defaults:
      run:
        shell: bash -el {0}
    steps:

    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        #repository: 'pytorch/pytorch'
        #ref: 'v2.3.0'
        submodules: 'recursive'

    - uses: conda-incubator/setup-miniconda@v3
      with:
        auto-activate-base: true
        activate-environment: true
        python-version: 3.10.13

    - name: Install Dependencies - Common - Linux 2
      run: |
        conda info
        conda list
        conda install nomkl
        conda install astunparse numpy ninja pyyaml setuptools cmake cffi typing_extensions future six requests dataclasses
        export PYTORCH_CPU_CAPABILITY=cpu
        export ATEN_CPU_CAPABILITY_DEFAULT=cpu
        export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
        export ATEN_CPU_CAPABILITY=default
        export USE_NNPACK=0
        export MAX_JOBS=4
        export USE_CUDA=0
        export USE_ROCM=0
        export BLAS=OpenBLAS
        export CMAKE_ARGS="-D CMAKE_BUILD_TYPE=Release -D USE_AVX=OFF -D USE_NNPACK=OFF -D C_HAS_AVX_2=OFF -D C_HAS_AVX2_2=OFF -D CXX_HAS_AVX_2=OFF -D CXX_HAS_AVX2_2=OFF -D CAFFE2_COMPILER_SUPPORTS_AVX512_EXTENSIONS=OFF -DPYTHON_INCLUDE_DIR=$(python -c "import sysconfig; print(sysconfig.get_path('include'))") -DPYTHON_LIBRARY=$(python -c "import sysconfig; print(sysconfig.get_config_var('LIBDIR'))") -DPYTHON_EXECUTABLE:FILEPATH=`which python`"
        pip install build wheel typing_extensions
        python setup.py bdist_wheel
    - name: Archive production artifacts
      uses: actions/upload-artifact@v4
      with:
        name: dist-without-markdown
        path: |
          dist
          !dist/**/*.md
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/126200
Approved by: https://github.com/jgong5, https://github.com/kit1980
 2024-05-31 19:38:42 +00:00
cyy
3e66052e16 Improve python3 discovery code in CMake (#127600) 
The improvement is based on my comments in #124613 and it also fixes the current linux-s390x-binary-manywheel  CI failures.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127600
Approved by: https://github.com/Skylion007
 2024-05-31 17:29:06 +00:00
cyy
0c5faee372 Replace python::python with Python::Module (#127485) 
Use found Python::Module target
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127485
Approved by: https://github.com/ezyang
 2024-05-31 05:57:05 +00:00
PyTorch MergeBot
7646825c3e Revert "distributed debug handlers (#126601)" 
This reverts commit 3d541835d509910fceca00fc5a916e9718c391d8.

Reverted https://github.com/pytorch/pytorch/pull/126601 on behalf of https://github.com/PaliC due to breaking internal typechecking tests ([comment](https://github.com/pytorch/pytorch/pull/126601#issuecomment-2141076987))
 2024-05-31 01:21:24 +00:00
cyy
d44daebdbc [Submodule] Remove deprecated USE_TBB option and TBB submodule (#127051) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127051
Approved by: https://github.com/cpuhrsch, https://github.com/malfet
 2024-05-31 01:20:45 +00:00
Tristan Rice
3d541835d5 distributed debug handlers (#126601) 
This adds debug handlers as described in:
* https://gist.github.com/d4l3k/828b7be585c7615e85b2c448b308d925 (public copy)
* https://docs.google.com/document/d/1la68szcS6wUYElUUX-P6zXgkPA8lnfzpagMTPys3aQ8/edit (internal copy)

This is only adding the C++ pieces that will be used from the main process. The Python and torchrun pieces will be added in a follow up PR.

This adds 2 handlers out of the box:

* `/handler/ping` for testing purposes
* `/handler/dump_nccl_trace_pickle` as a POC integration with Flight Recorder

Test plan:

```
python test/distributed/elastic/test_control_plane.py
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/126601
Approved by: https://github.com/kurman, https://github.com/c-p-i-o
 2024-05-30 02:21:08 +00:00
PyTorch MergeBot
67739d8c6f Revert "[Submodule] Remove deprecated USE_TBB option and TBB submodule (#127051)" 
This reverts commit 699db7988d84d163ebb6919f78885e4630182a7a.

Reverted https://github.com/pytorch/pytorch/pull/127051 on behalf of https://github.com/PaliC due to This PR needs to be synced using the import button as there is a bug in our diff train ([comment](https://github.com/pytorch/pytorch/pull/127051#issuecomment-2138496995))
 2024-05-30 01:16:57 +00:00
cyy
8ea1dc8748 Use Python::NumPy target (#127399) 
Now that we use FindPython, use it again for numpy detection.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127399
Approved by: https://github.com/malfet
 2024-05-29 23:17:58 +00:00
Nikita Shulga
0910429d72 [BE][CMake] Use FindPython module (#124613) 
As FindPythonInterp and FindPythonLibs has been deprecated since cmake-3.12

Replace `PYTHON_EXECUTABLE` with `Python_EXECUTABLE` everywhere (CMake variable names are case-sensitive)

This makes PyTorch buildable with python3 binary shipped with XCode on MacOS

TODO: Get rid of `FindNumpy` as its part of Python package
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124613
Approved by: https://github.com/cyyever, https://github.com/Skylion007
 2024-05-29 13:17:35 +00:00
cyy
699db7988d [Submodule] Remove deprecated USE_TBB option and TBB submodule (#127051) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127051
Approved by: https://github.com/cpuhrsch, https://github.com/malfet
 2024-05-29 11:58:03 +00:00
PyTorch MergeBot
cdbb2c9acc Revert "[Submodule] Remove deprecated USE_TBB option and TBB submodule (#127051)" 
This reverts commit 4fdbaa794f9d5af2f171f772a51cb710c51c925f.

Reverted https://github.com/pytorch/pytorch/pull/127051 on behalf of https://github.com/PaliC due to This PR needs to be synced using the import button as there is a bug in our diff train ([comment](https://github.com/pytorch/pytorch/pull/127051#issuecomment-2136428735))
 2024-05-29 03:02:35 +00:00
cyy
4fdbaa794f [Submodule] Remove deprecated USE_TBB option and TBB submodule (#127051) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127051
Approved by: https://github.com/cpuhrsch, https://github.com/malfet
 2024-05-27 03:54:03 +00:00
cyy
95e5c994f9 [Submodule] Clear USE_QNNPACK build option (#126941) 
Following the removal of QNNPACK third-party module #126657, we can clear more build system code. Also third_party/neon2sse was removed because it is not used.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/126941
Approved by: https://github.com/ezyang
 2024-05-24 00:12:56 +00:00