https://bugs.llvm.org/show_bug.cgi?id=43656
Bug ID: 43656
Summary: Regression in loop vectorizer (trunk vs. llvm 9)
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Loop Optimizer
Assignee: unassignedb...@nondot.org
Reporter: david.bolvan...@gmail.com
CC: llvm-bugs@lists.llvm.org
size_t h_dist (const uint8_t *restrict a, const uint8_t *restrict b, int n) {
size_t r = 0;
// #pragma clang loop vectorize(enable)
for (int i = 0; i < n; i++)
r += __builtin_popcount (a[i] ^ b[i]);
return r;
}
This loop is no longer vectorized with Clang trunk -O3 and we have a 20-30%
perf drop here.
The loop is vectorized with Clang <= 9.
remark: vectorized loop (vectorization width: 2, interleaved count: 2)
[-Rpass=loop-vectorize]
for (int i = 0; i < n; i++)
https://godbolt.org/z/Td8Ibo
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
