<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 - [X86] Add selective commutation support for insertps"
href="https://bugs.llvm.org/show_bug.cgi?id=40340">40340</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>[X86] Add selective commutation support for insertps
</td>
</tr>
<tr>
<th>Product</th>
<td>libraries
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>Windows NT
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>enhancement
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>Backend: X86
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>llvm-dev@redking.me.uk
</td>
</tr>
<tr>
<th>CC</th>
<td>craig.topper@gmail.com, llvm-bugs@lists.llvm.org, llvm-dev@redking.me.uk, spatel+llvm@rotateright.com
</td>
</tr></table>
<p>
<div>
<pre><a href="https://godbolt.org/z/bhohSK">https://godbolt.org/z/bhohSK</a>
#include <x86intrin.h>
__m128 ii(__m128 x, __m128 *y) {
return _mm_insert_ps(*y, x, (1<<6) | (1<<4) | 5);
}
define <4 x float> @_Z2iiDv4_fPS_(<4 x float>, <4 x float>* nocapture readonly)
{
%3 = load <4 x float>, <4 x float>* %1, align 16
%4 = tail call <4 x float> @llvm.x86.sse41.insertps(<4 x float> %3, <4 x
float> %0, i8 85)
ret <4 x float> %4
}
declare <4 x float> @llvm.x86.sse41.insertps(<4 x float>, <4 x float>, i8) #1
_Z2iiDv4_fPS_: # @_Z2iiDv4_fPS_
vmovaps (%rdi), %xmm1
vinsertps $85, %xmm0, %xmm1, %xmm0 # xmm0 = zero,xmm0[1],zero,xmm1[3]
retq
When we have 1 "inline" element from each source and zeros everywhere else we
should be able to commute the immediate to allow the memory fold.</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>