<html>
    <head>
      <base href="https://bugs.llvm.org/">
    </head>
    <body><table border="1" cellspacing="0" cellpadding="8">
        <tr>
          <th>Bug ID</th>
          <td><a class="bz_bug_link 
          bz_status_NEW "
   title="NEW - Compiler no longer optimizing some horizontal add/sub patterns after r310260"
   href="https://bugs.llvm.org/show_bug.cgi?id=34724">34724</a>
          </td>
        </tr>

        <tr>
          <th>Summary</th>
          <td>Compiler no longer optimizing some horizontal add/sub patterns after r310260
          </td>
        </tr>

        <tr>
          <th>Product</th>
          <td>new-bugs
          </td>
        </tr>

        <tr>
          <th>Version</th>
          <td>trunk
          </td>
        </tr>

        <tr>
          <th>Hardware</th>
          <td>PC
          </td>
        </tr>

        <tr>
          <th>OS</th>
          <td>Linux
          </td>
        </tr>

        <tr>
          <th>Status</th>
          <td>NEW
          </td>
        </tr>

        <tr>
          <th>Severity</th>
          <td>normal
          </td>
        </tr>

        <tr>
          <th>Priority</th>
          <td>P
          </td>
        </tr>

        <tr>
          <th>Component</th>
          <td>new bugs
          </td>
        </tr>

        <tr>
          <th>Assignee</th>
          <td>unassignedbugs@nondot.org
          </td>
        </tr>

        <tr>
          <th>Reporter</th>
          <td>douglas_yung@playstation.sony.com
          </td>
        </tr>

        <tr>
          <th>CC</th>
          <td>llvm-bugs@lists.llvm.org
          </td>
        </tr></table>
      <p>
        <div>
        <pre>Prior to r310260, the compiler was optimizing some add/sub + shuffle patterns
to horizontal add/sub instructions, but no longer seems to do so.

Consider this example:

__attribute__((noinline))
__m128 add_ps_001(__m128 a, __m128 b) {
  __m128 r = (__m128){ a[0] + a[1], a[2] + a[3], b[0] + b[1], b[2] + b[3] };
  return __builtin_shufflevector(r, a, -1, 1, 2, 3);
}

When compiled with the options "-S -O2 -march=bdver2" using a compiler prior to
r310260, the compiler would generate the following assembly:

vmovaps (%rcx), %xmm0
vhaddps (%rdx), %xmm0, %xmm0
retq

After r310260, the compiler is now generating the following code for the same
function:

vmovaps (%rdx), %xmm0
vmovddup        8(%rcx), %xmm1  # xmm1 = mem[0,0]
vinsertps       $28, %xmm0, %xmm1, %xmm2 # xmm2 = xmm1[0],xmm0[0],zero,zero
vinsertps       $76, %xmm1, %xmm0, %xmm1 # xmm1 = xmm1[1],xmm0[1],zero,zero
vaddps  %xmm1, %xmm2, %xmm1
vpermilpd       $1, %xmm0, %xmm2 # xmm2 = xmm0[1,0]
vpermilps       $231, %xmm0, %xmm0 # xmm0 = xmm0[3,1,2,3]
vaddss  %xmm0, %xmm2, %xmm0
vpermilps       $208, %xmm1, %xmm1 # xmm1 = xmm1[0,0,1,3]
vinsertps       $48, %xmm0, %xmm1, %xmm0 # xmm0 = xmm1[0,1,2],xmm0[0]
retq</pre>
        </div>
      </p>


      <hr>
      <span>You are receiving this mail because:</span>

      <ul>
          <li>You are on the CC list for the bug.</li>
      </ul>
    </body>
</html>