[MPS] Fix crash when max/min ops called for complex types by malfet · Pull Request #166214 · pytorch/pytorch

malfet · 2025-10-25T00:50:24Z

Stack from ghstack (oldest at bottom):

Raise an exception, as it's meaningless and results in segfault otherwise:

% python -c "import torch;torch.rand(10, dtype=torch.cfloat, device='mps').amax()"
(mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>'
(mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>>
(mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>'
(mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>>
/AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphExecutable.mm:1347: failed assertion `original module failed verification'
zsh: abort      python -c

To be tested by test_ops.py

[ghstack-poisoned]

pytorch-bot · 2025-10-25T00:50:28Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/166214

📄 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

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

✅ No Failures

As of commit b43a83f with merge base d049ed2 ():
💚 Looks good so far! There are no failures yet. 💚

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

[ghstack-poisoned]

malfet · 2025-10-31T02:35:41Z

@pytorchbot merge -f "Everything is green"

pytorchmergebot · 2025-10-31T02:37:12Z

Merge started

Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use -f as last resort and instead consider -i/--ignore-current to continue the merge ignoring current failures. This will allow currently pending tests to finish and report signal before the merge.

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

And enable fp16 implementation for CPU, which simplifies OpInfo definitions for the op Pull Request resolved: #166687 Approved by: https://github.com/Skylion007 ghstack dependencies: #166214

Or BatchNorm or LayerNorm for Long types Discovered while trying to enable `test_ops.py` for MPS Pull Request resolved: #166215 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: #166214, #166687

Raise an exception, as it's meaningless and results in segfault otherwise: ``` % python -c "import torch;torch.rand(10, dtype=torch.cfloat, device='mps').amax()" (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>' (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>> (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>' (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>> /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphExecutable.mm:1347: failed assertion `original module failed verification' zsh: abort python -c ``` To be tested by `test_ops.py` Pull Request resolved: #166214 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: #166272

And enable fp16 implementation for CPU, which simplifies OpInfo definitions for the op Pull Request resolved: #166687 Approved by: https://github.com/Skylion007 ghstack dependencies: #166214

Or BatchNorm or LayerNorm for Long types Discovered while trying to enable `test_ops.py` for MPS Pull Request resolved: #166215 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: #166214

…6214) Raise an exception, as it's meaningless and results in segfault otherwise: ``` % python -c "import torch;torch.rand(10, dtype=torch.cfloat, device='mps').amax()" (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>' (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>> (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: error: 'mps.reduction_max' op operand #0 must be tensor of mps native type values, but got 'tensor<10xcomplex<f32>>' (mpsFileLoc): /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm:176:0: note: see current operation: %2 = "mps.reduction_max"(%arg0, %1) <{keep_dims, propagate_nans}> : (tensor<10xcomplex<f32>>, tensor<1xsi32>) -> tensor<1xcomplex<f32>> /AppleInternal/Library/BuildRoots/4~B6shugDBannYeMBGCfhw7wjvNJOfy4BrawZ7TdI/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphExecutable.mm:1347: failed assertion `original module failed verification' zsh: abort python -c ``` To be tested by `test_ops.py` Pull Request resolved: pytorch#166214 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: pytorch#166272

And enable fp16 implementation for CPU, which simplifies OpInfo definitions for the op Pull Request resolved: pytorch#166687 Approved by: https://github.com/Skylion007 ghstack dependencies: pytorch#166214

Or BatchNorm or LayerNorm for Long types Discovered while trying to enable `test_ops.py` for MPS Pull Request resolved: pytorch#166215 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: pytorch#166214, pytorch#166687

Or BatchNorm or LayerNorm for Long types Discovered while trying to enable `test_ops.py` for MPS Pull Request resolved: #166215 Approved by: https://github.com/dcci, https://github.com/kulinseth, https://github.com/Skylion007 ghstack dependencies: #166214

Assert as it's meaningless otherwise ghstack-source-id: 2d9be90 Pull Request resolved: pytorch/pytorch#166214

Update

7f84339

[ghstack-poisoned]

pytorch-bot Bot added ciflow/mps Run MPS tests (subset of trunk) release notes: mps Release notes category labels Oct 25, 2025

This was referenced Oct 25, 2025

[BE][Opinfo] Mark [c]double as unsupported for MPS #166213

Closed

[MPS] Error out when BatchNorm is called for Complex #166215

Closed

malfet requested review from Skylion007 and dcci October 25, 2025 01:46

malfet added the topic: bug fixes topic category label Oct 25, 2025

dcci approved these changes Oct 25, 2025

View reviewed changes

malfet changed the title ~~[MPS] Fix crash when reduce ops called for scalar dtypes~~ [MPS] Fix crash when max/min ops called for complex types Oct 25, 2025

kulinseth approved these changes Oct 25, 2025

View reviewed changes

Skylion007 approved these changes Oct 26, 2025

View reviewed changes

Update

86e2e8c

[ghstack-poisoned]

This was referenced Oct 26, 2025

[MPS] Better error checking for FFT ops #166272

Closed

[MPS][1/N] Fix unsupported dtypes error checking for some MPS ops #166273

Closed

malfet and others added 3 commits October 26, 2025 18:46

Update

5a3f161

[ghstack-poisoned]

Update

5ae5ca3

[ghstack-poisoned]

Update

b43a83f

[ghstack-poisoned]

pytorchmergebot added the merging label Oct 31, 2025

pytorchmergebot closed this in 1257706 Oct 31, 2025

pytorchmergebot added Merged and removed merging labels Oct 31, 2025

Khanaksahu pushed a commit to Khanaksahu/pytorch that referenced this pull request Nov 17, 2025

[MPS] Fix crash when reduce ops called for complex dtypes

6db7612

Assert as it's meaningless otherwise ghstack-source-id: 2d9be90 Pull Request resolved: pytorch/pytorch#166214

Khanaksahu pushed a commit to Khanaksahu/pytorch that referenced this pull request Nov 17, 2025

[MPS] Fix crash when reduce ops called for complex dtypes

01e4129

Assert as it's meaningless otherwise ghstack-source-id: 2d9be90 Pull Request resolved: pytorch/pytorch#166214

github-actions Bot deleted the gh/malfet/571/head branch December 1, 2025 02:20

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[MPS] Fix crash when max/min ops called for complex types#166214

[MPS] Fix crash when max/min ops called for complex types#166214
malfet wants to merge 5 commits intogh/malfet/571/basefrom
gh/malfet/571/head

malfet commented Oct 25, 2025 •

edited

Loading

Uh oh!

pytorch-bot Bot commented Oct 25, 2025 •

edited

Loading

Uh oh!

malfet commented Oct 31, 2025

Uh oh!

pytorchmergebot commented Oct 31, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

malfet commented Oct 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot Bot commented Oct 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/166214

✅ No Failures

Uh oh!

malfet commented Oct 31, 2025

Uh oh!

pytorchmergebot commented Oct 31, 2025

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

malfet commented Oct 25, 2025 •

edited

Loading

pytorch-bot Bot commented Oct 25, 2025 •

edited

Loading