https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66580
Bug ID: 66580
Summary: max reduction does not auto vectorize
Product: gcc
Version: 5.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: dejan.crnila at dewesoft dot si
Target Milestone: ---
following code does (max reduction) not auto vectorize on x64/mingw gcc
#include <stdio.h>
#include <float.h>
int main()
{
float a[] __attribute__((aligned(0x20))) = {0, 1, 2, 3, 4, 5, 6, 7, 8,
9, 10, 11, 12, 13, 14, 15, 16};
float maxval = -FLT_MAX;
for (size_t i = 0; i < 16; i++)
maxval = a[i] > maxval ? a[i] : maxval;
printf("%g\n", maxval);
}
compiled as
gcc test.c -O3 -Wall -o test.exe -ftree-vectorizer-verbose=6
-fopt-info-vec-all -ffast-math -mavx
not supported or bug?
