Use atomic<int64_t> instead of atomic_int64_t. The later is available in GCC7, but not GCC5 #528
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This pull request was exported from Phabricator. Differential Revision: D34190739 |
…++14. Summary: std::shared_mutex is not available in C++14, and shows an error when incorporated into PyTorch CI. This patch replaces it with shared_timed_mutex, which is available, and is shared_mutex in GCC5.0. To use std::shared_mutex, we must use C++17. Differential Revision: D34190117 fbshipit-source-id: dab9bc72f5df9f050814e5e30ab73450b5995961
… in GCC7, but not GCC5 (pytorch#528) Summary: Pull Request resolved: pytorch#528 Since PyTorch CI builds for GCC5, we need to use std::atomic<std::int64_t> for iterationCount, because atomic_int64_t is not available until after GCC5. Test Plan: This fixes PyTorch CI error: https://app.circleci.com/pipelines/github/pytorch/pytorch/450702/workflows/e2f30e30-26d5-410d-8334-a2d6919d74de/jobs/16978938 ``` In file included from ../third_party/kineto/libkineto/src/ActivityProfilerController.cpp:3:0: ../third_party/kineto/libkineto/src/ActivityProfilerController.h:79:8: error: ‘atomic_int64_t’ in namespace ‘std’ does not name a type std::atomic_int64_t iterationCount_{-1}; ^ ../third_party/kineto/libkineto/src/ActivityProfilerController.cpp: In member function ‘void libkineto::ActivityProfilerController::step()’: ../third_party/kineto/libkineto/src/ActivityProfilerController.cpp:137:27: error: ‘iterationCount_’ was not declared in this scope int64_t currentIter = ++iterationCount_; ^ ../third_party/kineto/libkineto/src/ActivityProfilerController.cpp: In member function ‘void libkineto::ActivityProfilerController::scheduleTrace(const libkineto::Config&)’: ../third_party/kineto/libkineto/src/ActivityProfilerController.cpp:185:44: error: ‘iterationCount_’ was not declared in this scope if (config.hasProfileStartIteration() && iterationCount_ < 0) { ^ ``` Reviewed By: briancoutinho Differential Revision: D34190739 Pulled By: aaronenyeshi fbshipit-source-id: 6797ed39084a518d210a27cae2e6030cf8625c99
aaronenyeshi
force-pushed
the
export-D34190739
branch
from
24209d6 to
305d59a
Compare
|
This pull request was exported from Phabricator. Differential Revision: D34190739 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary: Since PyTorch CI builds for GCC5, we need to use std::atomicstd::int64_t for iterationCount, because atomic_int64_t is not available until after GCC5.
Reviewed By: briancoutinho
Differential Revision: D34190739
Pulled By: aaronenyeshi