[Bug tree-optimization/50693] Loop optimization restricted by GOTOs

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50693

--- Comment #10 from David Edelsohn <dje at gcc dot gnu.org> 2011-10-11 
01:35:20 UTC ---
Sorry, I was looking at the loop1 and loop2 functions, not the code inlined
into the benchmark for main.

LLVM generates:

        movq    %r12, %rdi
        movl    $99, %esi
        movq    %rbx, %rdx
        callq   memset

GCC vectorizes loop1:

.L22:
        addq    $1, %rdx
        movdqa  %xmm0, (%rcx)
        addq    $16, %rcx
        cmpq    %rsi, %rdx
        jb      .L22

but not loop2:

.L28:
.L29:
        movb    $99, (%rbx,%rax)
        addq    $1, %rax
        cmpq    %rbp, %rax
        jne     .L28