https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118213
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Version|unknown |15.0
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
You need to dive into IVOPTs details why it thinks this is better - it works
fine on x86_64:
.L3:
movzbl (%rsi,%rax), %r8d
cmpb %r8b, (%rdi,%rax)
jne .L6
addq $1, %rax
.L2:
cmpl %eax, %ecx
jne .L3
