https://llvm.org/bugs/show_bug.cgi?id=27619
Bug ID: 27619
Summary: [x86, loop vectorizer] cost model causes loop to not
vectorize
Product: libraries
Version: trunk
Hardware: PC
OS: All
Status: NEW
Severity: normal
Priority: P
Component: Loop Optimizer
Assignee: unassignedb...@nondot.org
Reporter: spatel+l...@rotateright.com
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
Re-filing the example from bug 27078 as a potential x86 cost model bug and/or
loop vectorizer bug. Note that this vectorizes for PPC but not x86 w/ AVX. But
I don't think the vectorization is being done optimally for PPC: no shuffles
should be needed if we treat everything as <12 x float>.
$ ./clang -O2 27078_orig.c -S -o - -emit-llvm -Rpass-analysis=loop-vectorize
-mavx
27078_orig.c:5:30: remark: the cost-model indicates that vectorization is not
beneficial [-Rpass-analysis=loop-vectorize]
float* blurred_pixel = blurred_row[j];
^
27078_orig.c:5:30: remark: the cost-model indicates that interleaving is not
beneficial [-Rpass-analysis=loop-vectorize]
typedef float Vector3_f[3];
void foo(Vector3_f* blurred_row, int width, float* pixel, float pixel_diff_avg)
{
for (int j = 0; j < width; ++j, pixel += 3) {
float* blurred_pixel = blurred_row[j];
float pixel_diff[3];
pixel_diff[0] = blurred_pixel[0] - pixel[0];
pixel_diff[1] = blurred_pixel[1] - pixel[1];
pixel_diff[2] = blurred_pixel[2] - pixel[2];
pixel_diff[0] -= pixel_diff_avg;
pixel_diff[1] -= pixel_diff_avg;
pixel_diff[2] -= pixel_diff_avg;
pixel[0] += pixel_diff[0];
pixel[1] += pixel_diff[1];
pixel[2] += pixel_diff[2];
}
}
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
