Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[ROCm] AMP selection process for AMD GPUs #47650

Merged
merged 7 commits into from
Apr 1, 2021
Merged

[ROCm] AMP selection process for AMD GPUs #47650

merged 7 commits into from
Apr 1, 2021

Conversation

stevenireeves
Copy link
Contributor

@stevenireeves stevenireeves commented Mar 8, 2021 •
edited

Chooses to enable based on AMD GPU architecture, with current supported arches being gfx906 and gfx908.
Utilizes the same Auto Mixed Precision tests as CUDA implementation.

@cheshire @chsigg for review.

@stevenireeves
Automatic Mixed Precision selection process for AMD GPUs based on arc…
e07a7d7 
…hitecture. Utilizes the same Auto Mixed Precision tests as CUDA.
@google-ml-butler google-ml-butler bot added the size:S CL Change Size: Small label Mar 8, 2021
@google-cla google-cla bot added the cla: yes label Mar 8, 2021
cheshire
cheshire requested changes Mar 9, 2021
View reviewed changes
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision.cc Outdated Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision_test.cc Show resolved Hide resolved
tensorflow/core/grappler/optimizers/auto_mixed_precision_test.cc Outdated Show resolved Hide resolved
@cheshire cheshire requested a review from reedwm March 9, 2021 01:02
@cheshire
Copy link
Member

cheshire commented Mar 9, 2021

Also adding Reed for sanity check on auto-mixed-precision stuff.

@gbaned gbaned self-assigned this Mar 9, 2021
@gbaned gbaned added the comp:gpu GPU related issues label Mar 9, 2021
@gbaned gbaned added this to Assigned Reviewer in PR Queue via automation Mar 9, 2021
@google-ml-butler google-ml-butler bot added kokoro:force-run Tests on submitted change ready to pull PR ready for merge process labels Mar 10, 2021
PR Queue automation moved this from Assigned Reviewer to Approved by Reviewer Mar 10, 2021
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 10, 2021
@google-ml-butler google-ml-butler bot removed the ready to pull PR ready for merge process label Mar 10, 2021
@stevenireeves
Copy link
Contributor Author

@cheshire I had a small mistake, which I fixed. Can we rerun the checks?

@google-ml-butler google-ml-butler bot added kokoro:force-run Tests on submitted change ready to pull PR ready for merge process labels Mar 10, 2021
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 10, 2021
@google-ml-butler google-ml-butler bot removed the ready to pull PR ready for merge process label Mar 10, 2021
@stevenireeves
Copy link
Contributor Author

@cheshire I diagnosed and fixed the Linux GPU failures given on commit b44d56b, tested locally. Everything should be ready.

PR Queue automation moved this from Approved by Reviewer to Reviewer Requested Changes Mar 10, 2021
@google-ml-butler google-ml-butler bot added kokoro:force-run Tests on submitted change ready to pull PR ready for merge process labels Mar 10, 2021
PR Queue automation moved this from Reviewer Requested Changes to Approved by Reviewer Mar 10, 2021
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 10, 2021
@google-ml-butler google-ml-butler bot added the kokoro:force-run Tests on submitted change label Mar 10, 2021
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 10, 2021
@google-ml-butler google-ml-butler bot removed the ready to pull PR ready for merge process label Mar 10, 2021
@stevenireeves
Copy link
Contributor Author

@cheshire @reedwm I accidentally pushed the wrong commit. The fix for the LinuxGPU failure was to move IsOnSupportedGPU after GetCudaVersion in auto_mixed_precision_test.cc. That is why auto_mixed_precision_test.cc couldn't compile the in the previous test. This should be the last review.

@google-ml-butler google-ml-butler bot added kokoro:force-run Tests on submitted change ready to pull PR ready for merge process labels Mar 10, 2021
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Mar 10, 2021
@gbaned gbaned added ready to pull PR ready for merge process and removed ready to pull PR ready for merge process labels Mar 12, 2021
@gbaned gbaned added ready to pull PR ready for merge process and removed ready to pull PR ready for merge process labels Mar 30, 2021
@copybara-service copybara-service bot merged commit c86acd4 into tensorflow:master Apr 1, 2021
PR Queue automation moved this from Approved by Reviewer to Merged Apr 1, 2021
@stevenireeves stevenireeves deleted the google_upstream_rocm_amp branch April 7, 2021 02:56
This was referenced Apr 9, 2021
Merged
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@cheshire cheshire cheshire approved these changes

@reedwm reedwm reedwm approved these changes

Assignees

@gbaned gbaned

Labels
cla: yes comp:gpu GPU related issues ready to pull PR ready for merge process size:S CL Change Size: Small
Projects
PR Queue
  
Merged
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
5 participants
@stevenireeves @cheshire @reedwm @kokoro-team @gbaned