conrad-watt opened PR #4113 from main to main:
This commit of the interpreter introduces a new strategy for handling bitwise ops, which fixes the performance of
popcnt,reinterpret, and signed integer comparison, and should result in fewer fuzzing timeouts (see https://github.com/bytecodealliance/wasmtime/issues/4059).For reference, similar performance issues still exist for the integer operations
clz,ctz,rem_s,rotl,rotr,wrap,extend_u, andextend_s- my current plan is to address these in a follow-up, so long as this current fix causes no issues.
conrad-watt edited PR #4113 from main to main:
This commit of the interpreter introduces a new strategy for handling bitwise ops, which fixes the performance of
popcnt,reinterpret, and signed integer comparison, and should result in fewer fuzzing timeouts (see https://github.com/bytecodealliance/wasmtime/issues/4059).For reference, similar performance issues still exist for the integer operations
clz,ctz,rem_s,rotl,rotr,wrap,extend_u, andextend_s- my current plan is to address these in a follow-up, so long as this current fix causes no issues.
conrad-watt edited PR #4113 from main to main:
This commit of the interpreter introduces a new strategy for handling bitwise ops, which fixes the performance of
popcnt,reinterpret, and signed integer comparison, and should result in fewer fuzzing timeouts (see https://github.com/bytecodealliance/wasmtime/issues/4059).For reference, similar performance issues still exist for the integer operations
clz,ctz,rem_s,rotl,rotr,wrap,extend_u, andextend_s- my current plan is to address these in a follow-up, so long as this current fix causes no issues.EDIT: currently fuzzing locally with no issues found so far
alexcrichton submitted PR review.
alexcrichton merged PR #4113.
Last updated: Dec 06 2025 at 06:05 UTC