https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112685
Bug ID: 112685
Summary: missed-optimization: division / modulo loops
Product: gcc
Version: 13.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: goon.pri.low at gmail dot com
Target Milestone: ---
This code here:
unsigned ldiv(unsigned r) {
unsigned d = 0;
while (r >= 3) {
r -= 3;
d++;
}
return d;
}
Could be optimized to a single division and save decades of processing time.
This code here:
unsigned lmod(unsigned r) {
while (r >= 3) r -= 3;
return r;
}
Could be optimized to a single modulo.
Signed integers are a whole other problem, but maybe we could optimize loops
for them also.
