afonso360 opened PR #6602 from afonso360:riscv-simd-extra-mul
to bytecodealliance:main
:
:wave: Hey,
This PR Implements
sqmul_round_sat
, we have a dedicated instruction for this so it's fairly straight forward.Additionally it also adds some variants of the existing multiplication instructions that splat one of the inputs. These were originally added before we had support for VX opcodes, and I've just noticed they were missing.
afonso360 requested fitzgen for a review on PR #6602.
afonso360 requested wasmtime-compiler-reviewers for a review on PR #6602.
afonso360 requested wasmtime-default-reviewers for a review on PR #6602.
afonso360 edited PR #6602:
:wave: Hey,
This PR Implements
sqmul_round_sat
, we have a dedicated instruction for this so it's fairly straight forward.Additionally it also adds some variants of the existing multiplication instructions that splat one of the inputs. These were originally added before we had support for VX opcodes, and I've just noticed they were missing.
elliottt submitted PR review:
This looks good to me! The
vx
variants for some of the vector ops seem extremely useful!
afonso360 merged PR #6602.
Last updated: Dec 23 2024 at 12:05 UTC