[Bug target/82447] New: Consider removing cmp instruction while iterating on an array of known bound

[antoshkka at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82447

            Bug ID: 82447
           Summary: Consider removing cmp instruction while iterating on
                    an array of known bound
           Product: gcc
           Version: 8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: antoshkka at gmail dot com
  Target Milestone: ---

Following code 

unsigned loop_read(unsigned char* a) {
    const unsigned size = 128;

    unsigned sum = 0;
    for (unsigned i = 0; i < size; ++i) {
        sum += a[i];
    }

    return sum;
}

Generates assembly:

loop_read(unsigned char*):
  lea rcx, [rdi+128]
  xor eax, eax
.L2:
  movzx edx, BYTE PTR [rdi]
  add rdi, 1
  add eax, edx
  cmp rdi, rcx  <=== This could be avoided
  jne .L2
  rep ret


The trick is to iterate from -128 to 0 and calling "jne .L2" right after the
increment. Here's how Clang does that:

loop_read(unsigned char*): # @loop_read(unsigned char*)
  xor eax, eax
  mov rcx, -128
.LBB0_1: # =>This Inner Loop Header: Depth=1
  movzx edx, byte ptr [rdi + rcx + 128]
  add eax, edx
  add rcx, 1
  jne .LBB0_1
  ret