[llvm-bugs] [Bug 33498] New: Can't coalesce four consecutive byte loads into a single x86 mov

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=33498

            Bug ID: 33498
           Summary: Can't coalesce four consecutive byte loads into a
                    single x86 mov
           Product: libraries
           Version: trunk
          Hardware: All
                OS: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: justin.lebar at gmail.com
                CC: llvm-bugs at lists.llvm.org, mkuper at google.com

Consider

unsigned foo(unsigned char* d) {
  return (static_cast<unsigned int>(d[0]) << 0) |
         (static_cast<unsigned int>(d[1]) << 8) |
         (static_cast<unsigned int>(d[2]) << 16) |
         (static_cast<unsigned int>(d[3]) << 24);
}

gcc 5.1+ compiles this as 

  mov     eax, DWORD PTR [rdi] 
  ret

but clang 4.0 -O2 naively does a series of shifts bitwise ORs:

        movzx   eax, byte ptr [rdi]
        movzx   ecx, byte ptr [rdi + 1]
        shl     ecx, 8
        or      ecx, eax
        movzx   edx, byte ptr [rdi + 2]
        shl     edx, 16
        or      edx, ecx
        movzx   eax, byte ptr [rdi + 3]
        shl     eax, 24
        or      eax, edx
        ret

I'll test on clang at HEAD in a few minutes once I get a build up and running.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20170618/c11f7ebb/attachment.html>