saymrwulf/pytorch
1
0
Fork 0
mirror of https://github.com/saymrwulf/pytorch.git synced 2026-05-14 20:57:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
83826 commits 5021 branches 1206 tags 1 GiB
Commit graph

1041 commits

Author SHA1 Message Date
Natalia Gimelshein
08ff11e9d0 initialize device when pinning memory on this device, short circuit i… (#145752) 
…s_pinned if device is not initialized
Do not land
RFC
potential fix for #144687

Now `.is_pinned(device="cuda")` does not initialize device and thus doesn't poison the fork (but it complains about `device` arg being deprecated). To not need `device=` arg we'd need to fix get_accelerator to not initialize device.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/145752
Approved by: https://github.com/albanD

Co-authored-by: albanD <albandes@fb.com>
 2025-01-30 21:37:29 +00:00
cyy
116af809eb Use std::string_view (#145906) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/145906
Approved by: https://github.com/albanD
 2025-01-30 03:14:27 +00:00
cyy
67fcc7cf02 [3/N] Remove unnecessary once flag usage (#145672) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/145672
Approved by: https://github.com/albanD
 2025-01-28 04:28:18 +00:00
Isuru Fernando
4b77ff9784 Fix PythonMod printing for C++ (#143385) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143385
Approved by: https://github.com/leslie-fang-intel, https://github.com/anijain2305
 2025-01-22 14:58:35 +00:00
Scott Wolchok
69b883d7ac Remove C10_EMBEDDED (#144808) 
I added this to support code sharing with ExecuTorch, but the operator<< overrides are load-bearing for builds -- we have other code that attempts to pretty-print Half/BFloat16, and implicit conversions can't be used to make that work because there are *multiple* implicit conversions from Half/BFloat16 to primitive types, so which one to select is ambiguous. Also, we don't actually seem to need it now in ExecuTorch core because we have `include <ostream>` in there at the moment anyway.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/144808
Approved by: https://github.com/janeyx99, https://github.com/malfet
 2025-01-15 06:08:53 +00:00
Richard Barnes
1dab79470d c10::string_view -> std::string_view in pytorch (#143591) 
Test Plan: Sandcastle

Differential Revision: D67312322

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143591
Approved by: https://github.com/malfet
 2025-01-13 21:44:05 +00:00
Nikita Shulga
9ae35b8bb1 [BE] Introduce c10::SyntaxError (#144647) 
Which will be translated into Python's SyntaxError
Pull Request resolved: https://github.com/pytorch/pytorch/pull/144647
Approved by: https://github.com/Skylion007
 2025-01-12 23:23:54 +00:00
Scott Wolchok
0529908f13 Remove is_reduced_floating_point from namespace std (#144502) 
Partial fix for #144495. Avoiding BC-break using existing practice of removing only if FBCODE_CAFFE2 and C10_NODEPRECATED are not defined.

Differential Revision: [D67992342](https://our.internmc.facebook.com/intern/diff/D67992342/)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/144502
Approved by: https://github.com/malfet
 2025-01-10 03:24:10 +00:00
cyy
9a841f9321 Enable bugprone-unchecked-optional-access (#144226) 
We can actually enable bugprone-unchecked-optional-access without the risk of hang.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/144226
Approved by: https://github.com/albanD
 2025-01-10 03:16:56 +00:00
cyy
d0070ca07e [18/N] Fix extra warnings brought by clang-tidy-17 (#144014) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/144014
Approved by: https://github.com/Skylion007, https://github.com/albanD
 2025-01-08 17:21:55 +00:00
cyy
dca443835e Enable more readability-redundant checks (#143963) 
They are helpful to simplifying code.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143963
Approved by: https://github.com/albanD
 2024-12-30 14:49:33 +00:00
Richard Barnes
e2d47a133b Disable c10::optional macros (#138912) 
Test Plan: Sandcastle

Pull Request resolved: https://github.com/pytorch/pytorch/pull/138912
Approved by: https://github.com/Skylion007, https://github.com/malfet
 2024-12-17 09:22:47 +00:00
Richard Barnes
82ce888273 c10::string_view -> std::string_view in more places (#142517) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/142517
Approved by: https://github.com/malfet
 2024-12-12 19:45:59 +00:00
cyy
7c1d5db1f3 [2/N] Enable UBSAN tests (#141740) 
Apply c10::load in more places. The function was introduced to cast a byte to valid boolean values, thus fixing the UBSAN errors.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/141740
Approved by: https://github.com/ezyang
 2024-12-03 20:52:26 +00:00
Richard Barnes
fca0f34b83 Switch c10::string_view to std::string_view (#139635) 
Shortens `string_view_starts_with` to `starts_with`. Adds some missing headers. Isolates `c10_string_view` to use with `get_fully_qualified_name`.

Test Plan: Sandcastle

Reviewed By: ezyang

Differential Revision: D64833558

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139635
Approved by: https://github.com/Skylion007, https://github.com/ezyang
 2024-11-27 01:41:18 +00:00
Kiuk Chung
54f4621ca5 Add missing explicit include directive for <cerrno> in c10/util/error… (#141593) 
`c10/util/error.cpp` uses the symbol `errno` but is missing an explicit header include directive for `<cerrno>`.

cc) @malfet , @atalman
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141593
Approved by: https://github.com/Skylion007
 2024-11-27 00:00:23 +00:00
Rafael Auler
68be990519 Flag TORCH_SDT_SEMAPHORE as being name resovable (#141191) 
Summary:
Mirroring changes in D64604573, it appears this code in libcaffe2 is
mostly a copy of folly's one. Copy of the original diff summary:

This particular inline assembly use cannot be converted to a constraint template parameter (until llvm 18 / gcc 14), as there is no way (until those versions) to specify that a non-pic relocation is needed when compiling under pic.  This inline assembly requires a non-pic relocation because it is being written to the .notes section, which is non .text.

Differential Revision: D66038989

Pull Request resolved: https://github.com/pytorch/pytorch/pull/141191
Approved by: https://github.com/dcci
 2024-11-23 01:39:44 +00:00
Scott Wolchok
75bbad4768 Unbreak CUDA 11.4 build of Half.h (#141173) 
`__CUDACC__` is needed to detect CUDA builds on that platform.

Differential Revision: [D66262774](https://our.internmc.facebook.com/intern/diff/D66262774/)

**NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D66262774/)!
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141173
Approved by: https://github.com/Skylion007, https://github.com/malfet
 2024-11-21 16:36:38 +00:00
cyy
d91484509a [1/N] Apply bugprone-unchecked-optional-access (#140679) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140679
Approved by: https://github.com/ezyang
 2024-11-20 04:04:41 +00:00
Scott Wolchok
f30f43f594 Use std::bit_cast as c10::bit_cast if available (#141035) 
Make what we're doing as obvious as possible to the compiler.

Differential Revision: [D66108811](https://our.internmc.facebook.com/intern/diff/D66108811/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141035
Approved by: https://github.com/Skylion007, https://github.com/ezyang, https://github.com/malfet
ghstack dependencies: #140564, #140565, #140566, #140567, #140720, #140994
 2024-11-19 20:43:45 +00:00
Scott Wolchok
efed02b990 Fix Half X86_F16 CUDA build failure (#140994) 
It passed PyTorch CI, but internally we saw failures from this.

Differential Revision: [D66137897](https://our.internmc.facebook.com/intern/diff/D66137897/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140994
Approved by: https://github.com/malfet
ghstack dependencies: #140564, #140565, #140566, #140567, #140720
 2024-11-19 19:02:21 +00:00
cyy
00b3b61076 Add and use thread-safe strerror (#140472) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140472
Approved by: https://github.com/ezyang
 2024-11-19 04:24:17 +00:00
Scott Wolchok
17bb78a3d3 Port X86_F16 from executorch half to PyTorch half (#140720) 
This was added in https://github.com/pytorch/executorch/pull/1789 . I'm working on sharing Half.h with ExecuTorch, and this is a missing feature.

Differential Revision: [D65949409](https://our.internmc.facebook.com/intern/diff/D65949409/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140720
Approved by: https://github.com/malfet
ghstack dependencies: #140564, #140565, #140566, #140567
 2024-11-18 21:32:44 +00:00
Scott Wolchok
081c1687c8 Remove UB type punning from c10/util/floating_point_utils.h (#140567) 
Accessing the inactive member of a union is undefined behavior. Fortunately, we have c10::bit_cast.

Differential Revision: [D65888680](https://our.internmc.facebook.com/intern/diff/D65888680/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140567
Approved by: https://github.com/Skylion007, https://github.com/malfet
ghstack dependencies: #140564, #140565, #140566
 2024-11-18 18:41:34 +00:00
Scott Wolchok
f59ec98ceb Add C10_EMBEDDED to gate ostream usage in Half/BFloat16 (#140566) 
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off.

Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140566
Approved by: https://github.com/ezyang, https://github.com/malfet
ghstack dependencies: #140564, #140565
 2024-11-18 18:41:34 +00:00
Scott Wolchok
e429a3b72e Move complex<Half> from Half.h to complex.h (#140565) 
Executing on old TODO on the way to sharing Half.h with ExecuTorch.

Differential Revision: [D65888037](https://our.internmc.facebook.com/intern/diff/D65888037/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140565
Approved by: https://github.com/ezyang, https://github.com/malfet
ghstack dependencies: #140564
 2024-11-18 15:56:21 +00:00
Scott Wolchok
f630799587 move c10::overflows to its own header (#140564) 
Working on moving `complex<Half>` to complex.h instead of Half.h; this depends on complex and isn't used particularly widely.

Differential Revision: [D65888038](https://our.internmc.facebook.com/intern/diff/D65888038/)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140564
Approved by: https://github.com/ezyang, https://github.com/Skylion007, https://github.com/malfet
 2024-11-18 15:56:21 +00:00
Shivam Raikundalia
c1bf714d76 [Profiler] Fix ASAN Overflow Issues (#140441) 
Summary:
It seems like this issues is due to leftover cupti events during warmup staying persistent in the queue during profiling. These events start before our actual time window and therefore have a timestamp lower than our basetime. This makes the delta become negative which results in unsigned overflow. This then creates a large number which later gets sign added which creates the signed overflow.

Solution: If a raw timestamp is less than the base timestamp, just mark the process timestamp as -1 so we can mark these events as "to ignore". In Kineto, add a special case to ignore timestamps that are negative.

Test Plan: Test with ASAN

Differential Revision: D65835650

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140441
Approved by: https://github.com/davidberard98
 2024-11-13 21:30:32 +00:00
PyTorch MergeBot
a58a565819 Revert "[Environment Variable][6/N] Use thread-safe getenv functions (#140200)" 
This reverts commit 7d4f5f7508.

Reverted https://github.com/pytorch/pytorch/pull/140200 on behalf of https://github.com/ezyang due to One of these diffs had incorrect downstream optional handling, we must reaudit all of these diffs ([comment](https://github.com/pytorch/pytorch/pull/140200#issuecomment-2473956859))
 2024-11-13 15:33:23 +00:00
cyy
7d4f5f7508 [Environment Variable][6/N] Use thread-safe getenv functions (#140200) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140200
Approved by: https://github.com/ezyang
 2024-11-09 15:05:51 +00:00
cyy
ab55a99283 Use TORCH_DECLARE_XXX (#139952) 
Because those files use TORCH_API

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139952
Approved by: https://github.com/ezyang
 2024-11-09 04:56:28 +00:00
PyTorch MergeBot
07ad74635b Revert "[Reland] Use static_assert to detect get_type_index used in device code (#139966)" 
This reverts commit ca7fdfe4d2.

Reverted https://github.com/pytorch/pytorch/pull/139966 on behalf of https://github.com/malfet due to This approach will prevent one from using get_type_index from device code ([comment](https://github.com/pytorch/pytorch/pull/139966#issuecomment-2465701260))
 2024-11-08 20:32:43 +00:00
cyy
ca7fdfe4d2 [Reland] Use static_assert to detect get_type_index used in device code (#139966) 
#139173 was reverted due to an internal build break of using get_type_index in device code. This PR is created for ease of importing into META to further investigation.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139966
Approved by: https://github.com/malfet, https://github.com/huydhn

Co-authored-by: Nikita Shulga <2453524+malfet@users.noreply.github.com>
 2024-11-07 17:36:47 +00:00
cyy
d558c1a047 Enable cppcoreguidelines-special-member-functions (#139132) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139132
Approved by: https://github.com/sraikund16
 2024-11-06 13:42:20 +00:00
PyTorch MergeBot
10d7729333 Revert "Enable cppcoreguidelines-special-member-functions (#139132)" 
This reverts commit a9b4989c72.

Reverted https://github.com/pytorch/pytorch/pull/139132 on behalf of https://github.com/ZainRizvi due to Sorry but this fails on trunk. See inductor/test_mkldnn_pattern_matcher.py::TestPatternMatcher::test_smooth_quant_with_int_mm [GH job link](https://github.com/pytorch/pytorch/actions/runs/11699366379/job/32591132460) [HUD commit link](22e89ea2aa) ([comment](https://github.com/pytorch/pytorch/pull/139132#issuecomment-2459743145))
 2024-11-06 13:27:42 +00:00
cyy
a9b4989c72 Enable cppcoreguidelines-special-member-functions (#139132) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139132
Approved by: https://github.com/sraikund16
 2024-11-06 07:59:09 +00:00
Judicaël Clair
1c63612567 Fix & unit test for c10::ArrayRef constructed from user-defined types (#139758) 
Fixes #139391

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139758
Approved by: https://github.com/ezyang
 2024-11-06 04:23:05 +00:00
cyy
a2bc2e38f9 Use clang-tidy 17 (#139678) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139678
Approved by: https://github.com/Skylion007
 2024-11-05 16:00:25 +00:00
cyy
64d9ee88d7 [11/N] Fix extra warnings brought by clang-tidy-17 (#139599) 
Follows #139385
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139599
Approved by: https://github.com/sraikund16
 2024-11-04 23:57:41 +00:00
cyyever
8ace3e8023 Add sv starts/ends_with (#139261) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139261
Approved by: https://github.com/Skylion007

Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>
 2024-11-01 01:17:42 +00:00
PyTorch MergeBot
886579af99 Revert "Use static_assert to detect get_type_index used in device code (#139173)" 
This reverts commit d391ed3f4e.

Reverted https://github.com/pytorch/pytorch/pull/139173 on behalf of https://github.com/facebook-github-bot due to Diff reverted internally ([comment](https://github.com/pytorch/pytorch/pull/139173#issuecomment-2450695123))
 2024-10-31 19:50:19 +00:00
cyy
d391ed3f4e Use static_assert to detect get_type_index used in device code (#139173) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139173
Approved by: https://github.com/r-barnes, https://github.com/ezyang
 2024-10-31 00:06:53 +00:00
cyy
4e5f9afc7f Enable c10::sv and std::sv constexpr conversions (#139239) 
As a small step towards moving c10::sv to std::sv and this tiny change shouldn't break META builds.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139239
Approved by: https://github.com/malfet
 2024-10-30 03:57:47 +00:00
cyy
3907f36808 Turn some variables and functions into static (#136847) 
Re-check some files and mark variables and functions into static and fix other warnings.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/136847
Approved by: https://github.com/ezyang
 2024-10-29 17:01:56 +00:00
cyy
e201460f8a [2/N] Fix Wextra-semi warnings (#139142) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139142
Approved by: https://github.com/ezyang
 2024-10-29 08:14:37 +00:00
cyy
383d9e3de6 [4/N] Fix cppcoreguidelines-special-member-functions warnings (#139027) 
Follows #138796
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139027
Approved by: https://github.com/ezyang
 2024-10-29 00:18:18 +00:00
cyy
f4f0f2995d Fix Wextra-semi warnings (#139000) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/139000
Approved by: https://github.com/ezyang
 2024-10-28 21:48:51 +00:00
Richard Barnes
42994234a6 std::value/std::type -> std::_v/std::_t (#138746) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/138746
Approved by: https://github.com/cyyever, https://github.com/malfet
 2024-10-26 20:59:24 +00:00
Taras
e78c4ded48 Use the unicode variant of the Windows API (#47422) (#138605) 
Use the unicode variant of the Windows API in c10/util/Backtrace.cpp
- #47422

Pull Request resolved: https://github.com/pytorch/pytorch/pull/138605
Approved by: https://github.com/peterjc123, https://github.com/malfet
 2024-10-26 17:41:39 +00:00
Kiuk Chung
7ada814107 [c10/util] Add explicit include of <mutex> to c10/util/env.cpp (#138854) 
Add explicit include of `<mutex>` to `c10/util/env.cpp` since it has usages of `std::lock_guard` which is defined in the header `<mutex>`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/138854
Approved by: https://github.com/cyyever, https://github.com/Skylion007
 2024-10-26 00:16:05 +00:00