macovedj opened PR #13966 from macovedj:winch-aarch64-v128-shuffle to bytecodealliance:main:
more https://github.com/bytecodealliance/wasmtime/issues/9925
macovedj requested cfallin for a review on PR #13966.
macovedj requested wasmtime-compiler-reviewers for a review on PR #13966.
macovedj requested wasmtime-core-reviewers for a review on PR #13966.
:thumbs_up: cfallin submitted PR review:
This looks reasonable to me, but some of the lowerings here are fairly complex -- how close are we to enabling some of the Wasm spec suite's SIMD tests that cover these with input+output test vectors?
(Happy to merge and you can answer off the critical path but it'd be great to go depth-first on getting some of those tests online if we can)
cfallin added PR #13966 winch(aarch64): v128 shuffle, swizzle, and bitmask ops to the merge queue.
github-merge-queue[bot] removed PR #13966 winch(aarch64): v128 shuffle, swizzle, and bitmask ops from the merge queue.
macovedj updated PR #13966.
macovedj commented on PR #13966:
This looks reasonable to me, but some of the lowerings here are fairly complex -- how close are we to enabling some of the Wasm spec suite's SIMD tests that cover these with input+output test vectors?
(Happy to merge and you can answer off the critical path but it'd be great to go depth-first on getting some of those tests online if we can)
All that's left to implement is NaN canonicalization which I'm reviewing to get ready after this lands. After that I should be able to start ungating the tests
cfallin has enabled auto merge for PR #13966.
cfallin added PR #13966 winch(aarch64): v128 shuffle, swizzle, and bitmask ops to the merge queue.
:check: cfallin merged PR #13966.
cfallin removed PR #13966 winch(aarch64): v128 shuffle, swizzle, and bitmask ops from the merge queue.
Last updated: Jul 29 2026 at 05:03 UTC