Conversation
Implements Option 2 from Issue #122: explicit naming convention w{weight}a{act}_{out}/ for clearer kernel identification. Directory mapping: - gemm/fp8/bf16/ -> gemm/w8a16_bf16/ (FP8 weight, BF16 activation) - gemm/fp8/fp8/ -> gemm/w8a8_bf16/ (pure FP8) - gemm/nvf4/bf16/ -> gemm/w4a16_bf16/ (NVF4 weight) - gemv/bf16/bf16/nvf4* -> gemv/w4a16_bf16/ - gemv/bf16/bf16/fp8* -> gemv/w8a16_bf16/ - gemv/fp8/fp8/ -> gemv/w8a8_bf16/ - gemv/nvf4/nvf4/ -> gemv/w4a4_bf16/ Updated: - CMakeLists.txt with new paths - Include paths in source files - CLAUDE.md with new naming convention docs Build verified: SM 120a CUDA 13.1 SUCCESS 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Implements Option 2 from Issue #122: explicit naming convention
w{weight}a{act}_{out}/for clearer kernel identification.Directory Mapping
GEMM:
gemm/fp8/bf16/gemm/w8a16_bf16/gemm/fp8/fp8/gemm/w8a8_bf16/gemm/fp8/f32/gemm/w8a8_f32/gemm/nvf4/bf16/gemm/w4a16_bf16/gemm/bf16/bf16/gemm/bf16_bf16/gemm/f32/f32/gemm/f32_f32/gemm/int8/int8/gemm/int8_int8/gemm/int4/int4/gemm/int4_int4/GEMV:
gemv/bf16/bf16/nvf4*gemv/w4a16_bf16/gemv/bf16/bf16/fp8*gemv/w8a16_bf16/gemv/fp8/fp8/gemv/w8a8_bf16/gemv/nvf4/nvf4/gemv/w4a4_bf16/gemv/bf16/bf16/gemv/bf16_bf16/gemv/int4/int4/gemv/int4_int4/Changes
git mv(history preserved)Naming Convention
The new convention
w{weight}a{act}_{out}makes kernel purpose immediately clear:w8a16_bf16: FP8 weights (8-bit), BF16 activations (16-bit), BF16 outputw4a16_bf16: NVF4 weights (4-bit), BF16 activations, BF16 outputw8a8_bf16: FP8 weights, FP8 activations, BF16 outputbf16_bf16: BF16 weights, BF16 activations (no quantization)Test Plan
Closes #122
🤖 Generated with Claude Code