https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90599
Bug ID: 90599
Summary: Inefficient code for __builtin_memset
Product: gcc
Version: 10.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: hjl.tools at gmail dot com
Target Milestone: ---
[hjl@gnu-cfl-1 pieces-6]$ cat x.i
extern char *dst;
void
foo (int x)
{
__builtin_memset (dst, x, 12);
}
[hjl@gnu-cfl-1 pieces-6]$ make x.s
/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/xgcc
-B/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/ -O3
-march=skylake -S x.i
[hjl@gnu-cfl-1 pieces-6]$ cat x.s
.file "x.i"
.text
.p2align 4
.globl foo
.type foo, @function
foo:
.LFB0:
.cfi_startproc
movzbl %dil, %eax
movabsq $72340172838076673, %rcx
movzbl %dil, %edi
imulq %rcx, %rax <<< x has been broadcasted to RAX.
imull $16843009, %edi, %edi
movq dst(%rip), %rdx
movq %rax, (%rdx)
movl %edi, 8(%rdx) <<< We should just reuse RAX/EAX.
ret
.cfi_endproc
.LFE0:
.size foo, .-foo
.ident "GCC: (GNU) 10.0.0 20190522 (experimental)"
.section .note.GNU-stack,"",@progbits
[hjl@gnu-cfl-1 pieces-6]$
