http://llvm.org/bugs/show_bug.cgi?id=17795
Bug ID: 17795
Summary: Loops with xors aren't vectorized
Product: libraries
Version: trunk
Hardware: Macintosh
OS: MacOS X
Status: NEW
Severity: normal
Priority: P
Component: Backend: X86
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Classification: Unclassified
Here is an example, I would expect the loop to be converted into a PXOR
instruction, or at least have a few iterations unrolled:
/tmp $ pygmentize t.c
#include <stdio.h>
void f(unsigned char *restrict a, unsigned char *restrict b) {
for (size_t i = 0; i < 128; i++) {
b[i] = 0x36 ^ a[i];
}
}
/tmp $ clang -O3 -S t.c
/tmp $ clang --version
Apple LLVM version 5.0 (clang-500.2.79) (based on LLVM 3.3svn)
Target: x86_64-apple-darwin13.0.0
Thread model: posix
/tmp $ cat t.s
.section __TEXT,__text,regular,pure_instructions
.globl _f
.align 4, 0x90
_f: ## @f
.cfi_startproc
## BB#0:
pushq %rbp
Ltmp2:
.cfi_def_cfa_offset 16
Ltmp3:
.cfi_offset %rbp, -16
movq %rsp, %rbp
Ltmp4:
.cfi_def_cfa_register %rbp
xorl %eax, %eax
.align 4, 0x90
LBB0_1: ## =>This Inner Loop Header: Depth=1
movb (%rdi,%rax), %cl
xorb $54, %cl
movb %cl, (%rsi,%rax)
incq %rax
cmpq $128, %rax
jne LBB0_1
## BB#2:
popq %rbp
ret
.cfi_endproc
.subsections_via_symbols
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
LLVMbugs mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
