MarinPostma opened PR #10147 from MarinPostma:packed-integer-arithmetic to bytecodealliance:main:
implements the following instructions for winch x64:
i8x16.addi8x16.add_sat_ui8x16.add_sat_si16x8.addi16x8.add_sat_ui16x8.add_sat_si32x4.addi64x2.addi8x16.subi8x16.sub_sat_ui8x16.sub_sat_si16x8.subi16x8.sub_sat_ui16x8.sub_sat_si32x4.subi64x2.subi8x16.muli16x8.muli32x4.muli64x2.mul#8093
MarinPostma requested fitzgen for a review on PR #10147.
MarinPostma requested wasmtime-compiler-reviewers for a review on PR #10147.
MarinPostma requested wasmtime-core-reviewers for a review on PR #10147.
MarinPostma edited PR #10147:
implements the following instructions for winch x64:
i8x16.addi8x16.add_sat_ui8x16.add_sat_si16x8.addi16x8.add_sat_ui16x8.add_sat_si32x4.addi64x2.addi8x16.subi8x16.sub_sat_ui8x16.sub_sat_si16x8.subi16x8.sub_sat_ui16x8.sub_sat_si32x4.subi64x2.subi16x8.muli32x4.muli64x2.mul#8093
MarinPostma updated PR #10147.
MarinPostma submitted PR review.
MarinPostma created PR review comment:
naming-wise, I'm completely sure what to call those. Maybe
vector_addis more appropriate?
saulecabrera commented on PR #10147:
I can take this review.
saulecabrera requested saulecabrera for a review on PR #10147.
MarinPostma updated PR #10147.
MarinPostma updated PR #10147.
saulecabrera submitted PR review.
saulecabrera created PR review comment:
I don't think we should make this a hard requirement, given that our baseline is AVX.
Given that Intel suggests that there's no penalty on mixing AVX with AVX512 instructions , we could emit AVX512 if they are available, however in case they aren't we still need to emit a fallback to avoid bumping our baseline for this operation. For reference: https://github.com/bytecodealliance/wasmtime/blob/main/cranelift/codegen/src/isa/x64/lower.isle#L1121
saulecabrera created PR review comment:
I think the naming is fine here, as far as I can tell, we don't have any other
vectornaming convention.
MarinPostma updated PR #10147.
MarinPostma updated PR #10147.
MarinPostma updated PR #10147.
MarinPostma submitted PR review.
MarinPostma created PR review comment:
my bad, I misunderstood your DM about what we expected to support. I have ported cranelift fallback implementation.
MarinPostma updated PR #10147.
MarinPostma submitted PR review.
MarinPostma created PR review comment:
hold on I just foudn a bug
MarinPostma updated PR #10147.
MarinPostma edited PR review comment.
MarinPostma updated PR #10147.
MarinPostma submitted PR review.
MarinPostma created PR review comment:
Fixed
MarinPostma updated PR #10147.
MarinPostma updated PR #10147.
saulecabrera submitted PR review:
LGTM, thanks!
saulecabrera merged PR #10147.
Last updated: Dec 13 2025 at 19:03 UTC