[ROCM] Navi21 Enablement 5: Softmax kernels #73545
Closed
+5
−3
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
CI Flow Status⚛️ CI FlowRuleset - Version:
|
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 9c83c14 (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
This was referenced Feb 28, 2022
Contributor
Author
dagitses
added
the
triaged
ngimel
approved these changes
Mar 3, 2022
Contributor
|
@ngimel has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Contributor
Author
|
@ngimel Did something break here? The commit has not been merged by the bot. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 8, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73545 Reviewed By: jbschlosser Differential Revision: D34616171 Pulled By: ngimel fbshipit-source-id: d9b3a17de2457e33ddc5d9c817799a1c85826ccb
Contributor
|
Hey @micmelesse. |
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 9, 2022
Summary: This PR is a follow up to the following prs. pytorch/pytorch#69942 pytorch/pytorch#72682 pytorch/pytorch#72809 pytorch/pytorch#73543 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch/pytorch#73545 Reviewed By: jbschlosser Differential Revision: D34616171 Pulled By: ngimel fbshipit-source-id: d9b3a17de2457e33ddc5d9c817799a1c85826ccb (cherry picked from commit f54b12c642e4b33cf9bf27f93f628e147dc37ddc)
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 9, 2022
Summary: This PR is a follow up to the following prs. pytorch/pytorch#69942 pytorch/pytorch#72682 pytorch/pytorch#72809 pytorch/pytorch#73543 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch/pytorch#73545 Reviewed By: jbschlosser Differential Revision: D34616171 Pulled By: ngimel fbshipit-source-id: d9b3a17de2457e33ddc5d9c817799a1c85826ccb (cherry picked from commit f54b12c642e4b33cf9bf27f93f628e147dc37ddc)
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 21, 2022
This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73546 Approved by: https://github.com/osalpekar
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 22, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73546 Approved by: https://github.com/osalpekar Test Plan: contbuild & OSS CI, see https://hud.pytorch.org/commit/pytorch/pytorch/14a891f38eb205169062e126ba81b5c9ececfc44 Reviewed By: malfet Differential Revision: D35026052 fbshipit-source-id: 64fb14d39199ccf2dafdb7e63b5fe78da315abf5
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 25, 2022
This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73548 Approved by: https://github.com/ngimel
shahofblah
pushed a commit
that referenced
this pull request
Mar 25, 2022
This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73546 Approved by: https://github.com/osalpekar
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 26, 2022
This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 #73548 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73549 Approved by: https://github.com/malfet
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 29, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73548 Approved by: https://github.com/ngimel Test Plan: contbuild & OSS CI, see https://hud.pytorch.org/commit/pytorch/pytorch/cd929f403f1a5d0a4feb9ec5a6bc6fe918d39a6e Reviewed By: malfet Differential Revision: D35188054 fbshipit-source-id: 630b45ba6b4d5b1386fcc0f8c979f41924fe9651
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 30, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 #73548 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73549 Approved by: https://github.com/malfet Test Plan: contbuild & OSS CI, see https://hud.pytorch.org/commit/pytorch/pytorch/56e0537e4e4fa209f70f0a08e82856c92c465162 Reviewed By: malfet Differential Revision: D35188063 fbshipit-source-id: b625fcff4acfa892a638b3cedde6c2818e68cd47
facebook-github-bot
pushed a commit
that referenced
this pull request
Apr 7, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 #73548 #73549 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73550 Reviewed By: malfet Differential Revision: D35444958 Pulled By: ngimel fbshipit-source-id: c65f06d3227c23bb097a71fc6c86e3f884114e04
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 7, 2022
Summary: This PR is a follow up to the following prs. #69942 #72682 #72809 #73543 #73545 #73546 #73548 #73549 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: #73550 Reviewed By: malfet Differential Revision: D35444958 Pulled By: ngimel fbshipit-source-id: c65f06d3227c23bb097a71fc6c86e3f884114e04 (cherry picked from commit 7f3ba52)
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 11, 2022
Summary: This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73545 Reviewed By: jbschlosser Differential Revision: D34616171 Pulled By: ngimel fbshipit-source-id: d9b3a17de2457e33ddc5d9c817799a1c85826ccb (cherry picked from commit f54b12c)
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 11, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73546 Approved by: https://github.com/osalpekar
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 11, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73548 Approved by: https://github.com/ngimel
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 11, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 pytorch#73548 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73549 Approved by: https://github.com/malfet
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 11, 2022
) Summary: This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 pytorch#73548 pytorch#73549 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73550 Reviewed By: malfet Differential Revision: D35444958 Pulled By: ngimel fbshipit-source-id: c65f06d3227c23bb097a71fc6c86e3f884114e04 (cherry picked from commit 7f3ba52)
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 12, 2022
Summary: This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73545 Reviewed By: jbschlosser Differential Revision: D34616171 Pulled By: ngimel fbshipit-source-id: d9b3a17de2457e33ddc5d9c817799a1c85826ccb (cherry picked from commit f54b12c)
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 12, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73546 Approved by: https://github.com/osalpekar
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 12, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73548 Approved by: https://github.com/ngimel
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 12, 2022
This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 pytorch#73548 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73549 Approved by: https://github.com/malfet
jeffdaily
pushed a commit
to ROCm/pytorch
that referenced
this pull request
Jul 12, 2022
) Summary: This PR is a follow up to the following prs. pytorch#69942 pytorch#72682 pytorch#72809 pytorch#73543 pytorch#73545 pytorch#73546 pytorch#73548 pytorch#73549 We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant. Pull Request resolved: pytorch#73550 Reviewed By: malfet Differential Revision: D35444958 Pulled By: ngimel fbshipit-source-id: c65f06d3227c23bb097a71fc6c86e3f884114e04 (cherry picked from commit 7f3ba52)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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.
This PR is a follow up to the following prs.
#69942
#72682
#72809
#73543
We are adding support to Navi21 GPUs which have a warpsize of 32. We cannot rely on a constant so we have to dynamically look up the warpsize when launching the kernel on the host side. Inside device functions this is not needed and the compiler can correctly detect the correct warpsize to replace the C10_WARP_SIZE constant.