feat: add MmShapeBench — 50 shapes x 2 dtypes = 100 matmul tests#51
Merged
Conversation
factnn
force-pushed
the
feature/mm-shape-bench-clean
branch
3 times, most recently
from
934936e to
65facc8
Compare
- 100 test files (50 HF model shapes x f16/f32) under kernelgenbench/accuracy/ - Registration in kernel_list.py and dataset/__init__.py - generate_prompts.py: MmShapeBench dataset support with shape context - generate_kernel_and_verify.py: MmShapePromptBuilder for op name mapping Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…enchmarks Add resolve_op_func_name() to handle specialized operators (e.g. MmShapeBench) where the benchmark op name (mm_128x768_768x768_f32) differs from the actual function name in generated code (mm). The verifier now resolves names before checking function definitions, removing the need for case-by-case prompt hacks.
factnn
force-pushed
the
feature/mm-shape-bench-clean
branch
from
65facc8 to
d4259f7
Compare
added 7 commits
June 26, 2026 14:57
For specialized benchmarks like MmShapeBench, the generator returns the base op name (aten::mm) but the verifier expects the full specialized op name (aten::mm_128x4096_4096x4096_f32). Use api_names[idx] to map back to the correct file name.
…add use_gems alias
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
aten::mmkernel_list.pyanddataset/__init__.pyfeature/mm-shape-bench, no unrelated changesFiles changed
src/flagbench/accuracy/mm_shape_bench/— 100 tests +__init__.pysrc/flagbench/dataset/__init__.py— add MM_SHAPE exportssrc/flagbench/dataset/kernel_list.py— add MM_SHAPE_OPERATORS dict