Repository navigation
perf: non-saturating i16 adds in CPU kernels, narrow CUDA tail stripe - #11
Merged
Merged
Conversation
Non-saturating adds give 1.2-1.6x avx512 kernel throughput; the transposed-kernel threshold drops to 256 nt.
jayhesselberth
force-pushed
the
worktree-perf-nonsat-add
branch
from
October 17, 2026 13:39
2a976cb to
164929e
Compare
This was referenced Oct 17, 2026
jayhesselberth
added a commit
that referenced
this pull request
Oct 17, 2026
The last stripe takes the narrowest of 16, 8 or 4 rows that covers the read, as the int kernel's has since #11: 1.09x local / 1.11x semi-global at 130 nt on an A30, still 80 registers. GpuScorer reports local memory.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Non-saturating i16 adds raise avx512 kernel throughput 1.2-1.6x; the CUDA tail stripe is narrowed and the transposed threshold drops to 256 nt. fmt, clippy, 120 tests and gpu_parity pass.
🤖 Generated with Claude Code