cmd/compile: call memmove on amd64 for large OpMove #25317
Labels
compiler/runtime
Issues related to the Go compiler and/or runtime.
NeedsInvestigation
Someone must examine and confirm this is a valid issue and not a duplicate of an existing one.
Performance
Milestone
Currently we lower OpMove with large size into REP MOV.
Memmove has optional AVX&co codepath, so it should be faster than REP MOV, or at least as fast if REP MOV is optimal (call overhead is insignificant for large sizes). We should probably generate a call to memmove for really big OpMove.
The text was updated successfully, but these errors were encountered: