8000 [Cutlass] Integrate EVT into CUDACPPScheduling by mlazos · Pull Request #150906 · pytorch/pytorch · GitHub

[Cutlass] Integrate EVT into CUDACPPScheduling #150906

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

Closed

mlazos wants to merge 37 commits into gh/mlazos/24/base from gh/mlazos/24/head

Contributor

mlazos commented

•

Stack from ghstack (oldest at bottom):

Previously merged:

Allow epilogue nodes in cuda combined scheduling

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov


          [Cutlass] Integrate EVT into CUDACPPScheduling

23ef06e

Allow epilogue nodes in cuda combined scheduling

[ghstack-poisoned]

pytorch-bot bot added ciflow/inductor module: inductor labels

pytorch-bot bot commented

•

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/150906

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit f20f30b with merge base ac9fcd6 ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

This was referenced Apr 9, 2025

[Cutlass] Import cutlass python API for EVT #150344

Closed

[Cutlass] Codegen for EVT Epilogue #150345

Closed

[Cutlass] Integrate EVT codegen into 3x gemm template #150346

Closed

[Cutlass] Implement Epilogue Argument emitter #150903

Closed

[Cutlass] Implement EVT example tensor creation #150904

Closed

[Cutlass] Implement cutlass epilogue visitor python codegen #150905

Closed

[Cutlass] Changes to gemm template for EVT #150907

Closed

[Inductor] Fix cuda_kernel typing #150908

Closed

[Inductor] Fix cuda_template.py typing #150909

Closed

[DO NOT MERGE] Throwaway changes #150910

Closed

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

9b6513d

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: a0a1aa0
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

144b406

Allow epilogue nodes in cuda combined scheduling

[ghstack-poisoned]

mlazos added the release notes: inductor label

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

90b0030

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 2fb6476
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

613528b

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 5110af4
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

19ecaf4

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

026b971

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: e478fb9
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

094448c

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

1753f4b

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: c2e1200
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

b0ed2ed

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

19099b4

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 0b9f5d8
Pull Request resolved: #150906

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

e2b32ab

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 0b9f5d8
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

22b9474

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

7aaf951

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: eadedfc
Pull Request resolved: #150906


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

91ad3db

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

ce5a72d

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 06f94a7
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

b5a803a

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 06f94a7
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

ad7cc67

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 06f94a7
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

10f931b

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos mentioned this pull request

[Inductor] Fix typing in cuda_template.py #152390

Closed

eellison reviewed

View reviewed changes

Contributor

eellison left a comment

Some of the conditions here should have already been checked in the non backend specific schedulilng..

add test ?

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

torch/_inductor/codegen/cuda/cuda_cpp_scheduling.py Outdated Show resolved Hide resolved

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

551f91c

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: abfdeac
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

fa6a8c5

Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

wdvr added the ci-no-td label

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

371e4a8

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: ec0ca8d
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

b1708be

Previously merged:
* #151713
* #151405
* #150905
* #152306
* #152305


Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

8000

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

643fe02

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 9466a8c
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling

Contributor Author

mlazos commented

Some of the conditions here should have already been checked in the non backend specific schedulilng..

add test ?

I'm going to add tests in the next PR, it needs the next change to work e2e

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

02230c6

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 9466a8c
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling

cuda_cpp_scheduling updates


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

6bb62cf

Previously merged:
* #151713
* #151405
* #150905
* #152306
* #152305


Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos mentioned this pull request

[Cutlass] Handle broadcasting in EVT python codegen #152733

Closed

mlazos requested a review from eellison

May 2, 2025 21:01

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

16a69a7

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: f264618
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling

cuda_cpp_scheduling updates


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

df5cd91

Previously merged:
* #151713
* #151405
* #150905
* #152306
* #152305


Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

mlazos added the ciflow/trunk label

mlazos mentioned this pull request

[Cutlass] E2E Tests for EVT #152815

Closed

mlazos added a commit that referenced this pull request


          [Cutlass] Integrate EVT into CUDACPPScheduling

b948286

Allow epilogue nodes in cuda combined scheduling

ghstack-source-id: 8e930e5
Pull Request resolved: #150906

Updates to cuda_cpp_scheduling

cuda_cpp_scheduling

cuda_cpp_scheduling updates


          Update on "[Cutlass] Integrate EVT into CUDACPPScheduling"

f20f30b

Previously merged:
* #151713
* #151405
* #150905
* #152306
* #152305


Allow epilogue nodes in cuda combined scheduling

cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy chenyang78 kadeng muchulee8 amjames chauhang aakhundov

[ghstack-poisoned]

eellison approved these changes

View reviewed changes

Collaborator

pytorchmergebot commented

Starting merge as part of PR stack under #151406

pytorchmergebot added the Merged label

pytorchmergebot closed this in

d483aef

pytorchmergebot pushed a commit that referenced this pull request


          [Cutlass] Add epilogue inputs/outputs to def_kernel (#151406)

df1ec04

Pull Request resolved: #151406
Approved by: https://github.com/eellison
ghstack dependencies: #152733, #150906

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ci-no-td ciflow/inductor ciflow/trunk Merged module: inductor release notes: inductor

< 2A4A /main>

0