[Inductor][CPP] Add Min/Max with VecMask #126841

leslie-fang-intel · 2024-05-22T03:08:46Z

Stack from ghstack (oldest at bottom):

Summary
Fix issue: #126824 which is missing the support of min/max with VecMask.

TestPlan

python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_max_cpu_bool
python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_min_cpu_bool

cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang

[ghstack-poisoned]

pytorch-bot · 2024-05-22T03:08:49Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/126841

📄 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.

✅ You can merge normally! (1 Unrelated Failure)

As of commit de9c5b9 with merge base 6079c50 ():

UNSTABLE - The following job failed but was likely due to flakiness present on trunk and has been marked as unstable:

inductor / rocm6.1-py3.8-inductor / test (inductor, 1, 1, linux.rocm.gpu.2) (gh) (#128871)
'Test'

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

cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 voznesenskym penguinwu EikanWang Guobing-Chen zhuhaozhe blzheng wenzhe-nrv jiayisunx peterbell10 ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang [ghstack-poisoned]

ghstack-source-id: f741652186bf5cf253d12fca40cc957aefd6b327 Pull Request resolved: #126841

torch/_inductor/codegen/cpp.py

**Summary** Fix issue: #126824 which is missing the support of `min/max` with `VecMask`. **TestPlan** ``` python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_max_cpu_bool python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_min_cpu_bool ``` cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 voznesenskym penguinwu EikanWang Guobing-Chen zhuhaozhe blzheng wenzhe-nrv jiayisunx peterbell10 ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang [ghstack-poisoned]

leslie-fang-intel · 2024-05-24T01:14:56Z

Hi @isuruf, these 2 PRs in the ghstack may have conflict with #124021. We prefer to land 124021 at first. Then I will rebase to land the PRs here. Does it sound good to you?

isuruf · 2024-05-24T03:35:01Z

Sounds good to me.

**Summary** Fix issue: #126824 which is missing the support of `min/max` with `VecMask`. **TestPlan** ``` python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_max_cpu_bool python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_min_cpu_bool ``` cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 voznesenskym penguinwu EikanWang Guobing-Chen zhuhaozhe blzheng wenzhe-nrv jiayisunx peterbell10 ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang [ghstack-poisoned]

ghstack-source-id: c0b307504bd97abc06c814b04c86882c382f4843 Pull Request resolved: pytorch#126841

leslie-fang-intel · 2024-05-25T23:45:51Z

@pytorchbot merge

**Summary** Following the discussion in pytorch#122593 (comment), Move Inductor MKLDNN specific IRs to a separate file. Pull Request resolved: pytorch#126504 Approved by: https://github.com/desertfire, https://github.com/jgong5 ghstack dependencies: pytorch#126841, pytorch#126940

This reverts commit bf2909b. Reverted pytorch#126504 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

This reverts commit f8c4c26. Reverted pytorch#126940 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

This reverts commit 1ef4306. Reverted pytorch#126841 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

This reverts commit bf2909b. Reverted pytorch#126504 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

This reverts commit f8c4c26. Reverted pytorch#126940 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

This reverts commit 1ef4306. Reverted pytorch#126841 on behalf of https://github.com/DanilBaibak due to Blocks reverting of the broken PR ([comment](pytorch#126841 (comment)))

isuruf · 2024-06-11T04:16:40Z

@pytorchbot rebase

pytorchmergebot · 2024-06-11T04:18:08Z

@pytorchbot started a rebase job onto refs/remotes/origin/viable/strict. Check the current status here

pytorchmergebot · 2024-06-11T04:18:15Z

Rebase failed due to

Aborting rebase because rebasing the branch resulted in the same sha as the target branch.
This usually happens because the PR has already been merged.  Please rebase locally and push.

Raised by https://github.com/pytorch/pytorch/actions/runs/9459495597

**Summary** Fix issue: #126824 which is missing the support of `min/max` with `VecMask`. **TestPlan** ``` python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_max_cpu_bool python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_clamp_min_cpu_bool ``` cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 voznesenskym penguinwu EikanWang Guobing-Chen zhuhaozhe blzheng wenzhe-nrv jiayisunx peterbell10 ipiszy yf225 chenyang78 kadeng muchulee8 ColinPeppler amjames desertfire chauhang [ghstack-poisoned]

isuruf · 2024-06-17T17:40:49Z

test/inductor/test_torchinductor_opinfo.py

- "clamp_max": {b8},
- "clamp_min": {b8},


Aside from these two, you can remove max.binary, min.binary, maximum, minimum too

Oh, just saw the comment. I will send a standalone PR to remove it.

leslie-fang-intel · 2024-06-18T00:18:44Z

@pytorchbot merge

pytorchmergebot · 2024-06-18T00:20:20Z

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

**Summary** Fix #126824 (comment) which is missing the support of `ne` with `VecMask`. **Test Plan** ``` python test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_ne_cpu_bool ``` Co-authored-by: Isuru Fernando <[email protected]> Pull Request resolved: #126940 Approved by: https://github.com/isuruf, https://github.com/jgong5, https://github.com/peterbell10 ghstack dependencies: #126841

**Summary** Following the discussion in #122593 (comment), Move Inductor MKLDNN specific IRs to a separate file. Co-authored-by: Isuru Fernando <[email protected]> Pull Request resolved: #126504 Approved by: https://github.com/desertfire, https://github.com/jgong5 ghstack dependencies: #126841, #126940

**Summary** Remove `max.binary, min.binary, maximum, minimum` from `inductor_one_sample` op list as we fix the bool vectorization issue in #126841. **Test Plan** ``` python -u -m pytest -s -v test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_maximum python -u -m pytest -s -v test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_minimum python -u -m pytest -s -v test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_min_binary python -u -m pytest -s -v test/inductor/test_torchinductor_opinfo.py -k test_comprehensive_max_binary ``` Pull Request resolved: #128925 Approved by: https://github.com/isuruf, https://github.com/jgong5, https://github.com/peterbell10

Add Min/Max with VecMask

1f53771

[ghstack-poisoned]

pytorch-bot bot added ciflow/inductor module: cpu CPU specific problem (e.g., perf, algorithm) module: inductor labels May 22, 2024

leslie-fang-intel added the ciflow/trunk Trigger trunk jobs on your pull request label May 22, 2024

leslie-fang-intel added a commit that referenced this pull request May 22, 2024

Add Min/Max with VecMask

6dbd978

ghstack-source-id: f741652186bf5cf253d12fca40cc957aefd6b327 Pull Request resolved: #126841

pytorchbot added the open source label May 22, 2024

leslie-fang-intel requested review from jgong5, peterbell10 and isuruf May 22, 2024 05:14

leslie-fang-intel changed the title ~~Add Min/Max with VecMask~~ [Inductor][CPP] Add Min/Max with VecMask May 22, 2024

isuruf approved these changes May 22, 2024

View reviewed changes

jgong5 approved these changes May 22, 2024

View reviewed changes

peterbell10 reviewed May 22, 2024

View reviewed changes

torch/_inductor/codegen/cpp.py Outdated Show resolved Hide resolved

peterbell10 approved these changes May 22, 2024

View reviewed changes

leslie-fang-intel mentioned this pull request May 23, 2024

[Inductor][CPP] Add ne with VecMask #126940

Closed

leslie-fang-intel mentioned this pull request May 24, 2024

Move MKLDNN Specific IR to Separate File #126504

Closed

leslie-fang-intel added 2 commits May 24, 2024 07:44

leslie-fang-intel added a commit to leslie-fang-intel/pytorch that referenced this pull request May 25, 2024

Add Min/Max with VecMask

bbdcf4e

ghstack-source-id: c0b307504bd97abc06c814b04c86882c382f4843 Pull Request resolved: pytorch#126841

isuruf and others added 4 commits June 12, 2024 20:05

isuruf reviewed Jun 17, 2024

View reviewed changes

pytorchmergebot added the merging label Jun 18, 2024

pytorchmergebot closed this in beb2983 Jun 18, 2024

pytorchmergebot removed the merging label Jun 18, 2024

leslie-fang-intel mentioned this pull request Jun 18, 2024

[Inductor] Remove min/max from inductor opinfo test #128925

Closed

github-actions bot deleted the gh/leslie-fang-intel/102/head branch July 18, 2024 01:56

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Inductor][CPP] Add Min/Max with VecMask #126841

[Inductor][CPP] Add Min/Max with VecMask #126841

leslie-fang-intel commented May 22, 2024 •

edited

Loading

pytorch-bot bot commented May 22, 2024 •

edited

Loading

leslie-fang-intel commented May 24, 2024

isuruf commented May 24, 2024

leslie-fang-intel commented May 25, 2024

isuruf commented Jun 11, 2024

pytorchmergebot commented Jun 11, 2024

pytorchmergebot commented Jun 11, 2024

isuruf Jun 17, 2024

leslie-fang-intel Jun 18, 2024

leslie-fang-intel commented Jun 18, 2024

pytorchmergebot commented Jun 18, 2024

[Inductor][CPP] Add Min/Max with VecMask #126841

[Inductor][CPP] Add Min/Max with VecMask #126841

Conversation

leslie-fang-intel commented May 22, 2024 • edited Loading

pytorch-bot bot commented May 22, 2024 • edited Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/126841

✅ You can merge normally! (1 Unrelated Failure)

leslie-fang-intel commented May 24, 2024

isuruf commented May 24, 2024

leslie-fang-intel commented May 25, 2024

isuruf commented Jun 11, 2024

pytorchmergebot commented Jun 11, 2024

pytorchmergebot commented Jun 11, 2024

isuruf Jun 17, 2024

Choose a reason for hiding this comment

leslie-fang-intel Jun 18, 2024

Choose a reason for hiding this comment

leslie-fang-intel commented Jun 18, 2024

pytorchmergebot commented Jun 18, 2024

Merge started

leslie-fang-intel commented May 22, 2024 •

edited

Loading

pytorch-bot bot commented May 22, 2024 •

edited

Loading