[ET-VK][ez] Fix NaN propagation in binary div operations for padded texels #16419
+26
−5
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
…exels When the packed dimension size is not a multiple of 4, texture-backed tensors have padding elements in the last texel. For division operations, these padding regions contain 0/0 = NaN, which propagates through subsequent reduction operations and corrupts results. This fix adds conditional padding masking logic to binary_op shaders: - Introduced MASK_PADDING codegen variable to binary_op.yaml - Enabled MASK_PADDING=1 for binary_div and binary_floor_divide ops - Added GLSL preprocessor macro definition in binary_op.glsl - Implemented padding masking logic using modulo arithmetic to correctly identify last texels in batch concatenation scenarios - Padding elements are explicitly zeroed out to prevent NaN propagation The implementation follows GLSL best practices by using Python preprocessing only for macro definition, keeping core shader logic as pure GLSL with standard #ifdef directives. Differential Revision: [D89935220](https://our.internmc.facebook.com/intern/diff/D89935220/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16419
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 2 New Failures, 1 Unrelated FailureAs of commit da3a2f1 with merge base 1066e7c ( NEW FAILURES - The following jobs have failed:
UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
This PR needs a
|
…or padded texels" When the packed dimension size is not a multiple of 4, texture-backed tensors have padding elements in the last texel. For division operations, these padding regions contain 0/0 = NaN, which propagates through subsequent reduction operations and corrupts results. This fix adds conditional padding masking logic to binary_op shaders: - Introduced MASK_PADDING codegen variable to binary_op.yaml - Enabled MASK_PADDING=1 for binary_div and binary_floor_divide ops - Added GLSL preprocessor macro definition in binary_op.glsl - Implemented padding masking logic using modulo arithmetic to correctly identify last texels in batch concatenation scenarios - Padding elements are explicitly zeroed out to prevent NaN propagation The implementation follows GLSL best practices by using Python preprocessing only for macro definition, keeping core shader logic as pure GLSL with standard #ifdef directives. Differential Revision: [D89935220](https://our.internmc.facebook.com/intern/diff/D89935220/) [ghstack-poisoned]
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Stack from ghstack (oldest at bottom):
When the packed dimension size is not a multiple of 4, texture-backed
tensors have padding elements in the last texel. For division operations,
these padding regions contain 0/0 = NaN, which propagates through
subsequent reduction operations and corrupts results.
This fix adds conditional padding masking logic to binary_op shaders:
identify last texels in batch concatenation scenarios
The implementation follows GLSL best practices by using Python
preprocessing only for macro definition, keeping core shader logic
as pure GLSL with standard #ifdef directives.
Differential Revision: D89935220