Closed MJoergen closed 3 years ago
I'm closing this Issue again.
I examined the generated assembly code for __lslint32
and __lsrint32
, and they are already highly optimized with an implementation close to that of the corresponding monitor calls.
The library functions
__lslint32
and__lsrint32
can be optimized by making use of the monitor callsshl32
andshr32
.