Skip to content

[ROCm] Adjust elementwise_kernel settings on ROCm#32609

Closed
iotamudelta wants to merge 5 commits intopytorch:masterfrom
iotamudelta:adjust_loop_bounds
Closed

[ROCm] Adjust elementwise_kernel settings on ROCm#32609
iotamudelta wants to merge 5 commits intopytorch:masterfrom
iotamudelta:adjust_loop_bounds

Conversation

@iotamudelta
Copy link
Contributor

Recent PR #31974 and upcoming PR #32383 are changing the behavior of the elementwise_kernel infrastructure on CUDA.

In order to stay in sync, change the nd-loop behavior to match ROCm and CUDA for now. Once the full rework is done, the ROCm settings will likely diverge again.

@iotamudelta iotamudelta added module: rocm AMD GPU support for Pytorch open source labels Jan 25, 2020
@kostmo
Copy link
Member

kostmo commented Jan 25, 2020

💊 CircleCI build failures summary and remediations

As of commit d64e167:

  • 1/1 failures introduced in this PR

Detailed failure analysis

One may explore the probable reasons each build failed interactively on the Dr. CI website.

🕵️ 1 new failure recognized by patterns

The following build failures do not appear to be due to upstream breakage:

See CircleCI build pytorch_xla_linux_xenial_py3_6_clang7_build (1/1)

Step: "Build" (full log | pattern match details)

Jan 25 03:58:46 RuntimeError: Failed to apply patch: /var/lib/jenkins/workspace/xla/torch_patches/X10-device_test.diff
Jan 25 03:58:46 File to patch:  
Jan 25 03:58:46 Skip this patch? [y]  
Jan 25 03:58:46 4 out of 4 hunks ignored 
Jan 25 03:58:46 Applying patch file: /var/lib/jenkins/workspace/xla/torch_patches/X10-clip_grad.diff 
Jan 25 03:58:46 Applying patch file: /var/lib/jenkins/workspace/xla/torch_patches/X10-device_test.diff 
Jan 25 03:58:46 Traceback (most recent call last): 
Jan 25 03:58:46   File "/var/lib/jenkins/workspace/xla/scripts/cond_patch.py", line 67, in <module> 
Jan 25 03:58:46     patch_repo(args) 
Jan 25 03:58:46   File "/var/lib/jenkins/workspace/xla/scripts/cond_patch.py", line 49, in patch_repo 
Jan 25 03:58:46     raise RuntimeError('Failed to apply patch: {}'.format(ppath)) 
Jan 25 03:58:46 RuntimeError: Failed to apply patch: /var/lib/jenkins/workspace/xla/torch_patches/X10-device_test.diff 
Jan 25 03:58:46 + cleanup 
Jan 25 03:58:46 + retcode=1 
Jan 25 03:58:46 + set +x 
Jan 25 03:58:46 =================== sccache compilation log =================== 
Jan 25 03:58:46 =========== If your build fails, please take a look at the log above for possible reasons =========== 
Jan 25 03:58:46 Compile requests                 1 
Jan 25 03:58:46 Compile requests executed        0 
Jan 25 03:58:46 Cache hits                       0 
Jan 25 03:58:46 Cache misses                     0 
Jan 25 03:58:46 Cache timeouts                   0 

This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker.

This comment has been revised 2 times.

@ngimel ngimel added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jan 25, 2020
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ezyang is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@facebook-github-bot
Copy link
Contributor

@ezyang merged this pull request in 5ac2593.

1 similar comment
@facebook-github-bot
Copy link
Contributor

@ezyang merged this pull request in 5ac2593.

wuhuikx pushed a commit to wuhuikx/pytorch that referenced this pull request Jan 30, 2020
Summary:
Recent PR pytorch#31974 and upcoming PR pytorch#32383 are changing the behavior of the elementwise_kernel infrastructure on CUDA.

In order to stay in sync, change the nd-loop behavior to match ROCm and CUDA for now. Once the full rework is done, the ROCm settings will likely diverge again.
Pull Request resolved: pytorch#32609

Differential Revision: D19580121

Pulled By: ezyang

fbshipit-source-id: 4c8dcf6db3ac973e48ece6a665615cfe7d7cb764
ttumiel pushed a commit to ttumiel/pytorch that referenced this pull request Mar 4, 2020
Summary:
Recent PR pytorch#31974 and upcoming PR pytorch#32383 are changing the behavior of the elementwise_kernel infrastructure on CUDA.

In order to stay in sync, change the nd-loop behavior to match ROCm and CUDA for now. Once the full rework is done, the ROCm settings will likely diverge again.
Pull Request resolved: pytorch#32609

Differential Revision: D19580121

Pulled By: ezyang

fbshipit-source-id: 4c8dcf6db3ac973e48ece6a665615cfe7d7cb764
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Merged module: rocm AMD GPU support for Pytorch open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants