[DEV][PROFILER] Case 3: Add mask usage percentage tracking to Profiler #192
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
Add CHECK_LOAD_MASK_PERCENTAGE flag to Profiler to track statistics on masked vs unmasked load/store operations. This helps identify optimization opportunities where masked operations could be replaced with more efficient unmasked variants. Changes: - Add counters for total/masked loads and stores in Profiler - Implement mask detection in load/store callbacks - Add finalize() output to display mask usage statistics - Refactor CHECK_BUFFER_LOAD into a configurable flag - Add comprehensive test to verify mask percentage counting
Jokeren
reviewed
Oct 16, 2025
| def __init__( | ||
| self, | ||
| callpath: bool = True, | ||
| CHECK_BUFFER_LOAD: bool = False, |
There was a problem hiding this comment.
I think it's fine to turn on all these options by default
| assert False, "Buffer Load optimization should be used when offsets are within 32-bit range!" | ||
|
|
||
| def register_op_callback(self, op_type: type[Op]) -> OpCallbacks: | ||
| def _is_mask_all_true(mask: TensorHandle) -> bool: |
There was a problem hiding this comment.
You misunderstood here. I wanted to check the percentage of true values in the mask, not "all true".
For example if the mask is [true true false false] then it's 50%
If we have two masks we sum up all trues and falses
mark14wu
changed the title
|
@Jokeren It's ready for review now. If no more issues, I can run a pass of case 3 checking through TritonBench. |
Jokeren
approved these changes
Oct 18, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
Add CHECK_LOAD_MASK_PERCENTAGE flag to Profiler to track statistics on masked vs unmasked load/store operations. This helps identify optimization opportunities where masked operations could be replaced with more efficient unmasked variants.
Changes: