<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 - [AVX512][Regression] Failed to use vpsubusw"
href="https://bugs.llvm.org/show_bug.cgi?id=48223">48223</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>[AVX512][Regression] Failed to use vpsubusw
</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>Linux
</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>david.bolvansky@gmail.com
</td>
</tr>
<tr>
<th>CC</th>
<td>craig.topper@gmail.com, llvm-bugs@lists.llvm.org, llvm-dev@redking.me.uk, pengfei.wang@intel.com, spatel+llvm@rotateright.com
</td>
</tr></table>
<p>
<div>
<pre>#define SIZE 32
#define WSIZE 64
unsigned short head[SIZE] __attribute__((aligned(64)));
void f(void)
{
for (unsigned n = 0; n < SIZE; ++n) {
unsigned short m = head[n];
head[n] = (unsigned short)(m >= WSIZE ? m-WSIZE : 0);
}
}
Clang 11 or newer with -O3 -mavx512f:
f(): # @f()
vmovdqa ymm0, ymmword ptr [rip + .LCPI0_0] # ymm0 =
[64,64,64,64,64,64,64,64,64,64,64,64,64,64,64,64]
vmovdqa ymm1, ymmword ptr [rip + head+32]
vpmaxuw ymm2, ymm1, ymm0
vpcmpeqw ymm2, ymm1, ymm2
vmovdqa ymm3, ymmword ptr [rip + head]
vpmaxuw ymm0, ymm3, ymm0
vpcmpeqw ymm0, ymm3, ymm0
vinserti64x4 zmm0, zmm0, ymm2, 1
vmovdqa ymm2, ymmword ptr [rip + .LCPI0_1] # ymm2 =
[65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472,65472]
vpaddw ymm1, ymm1, ymm2
vpaddw ymm2, ymm3, ymm2
vinserti64x4 zmm1, zmm2, ymm1, 1
vpandq zmm0, zmm0, zmm1
vmovdqa64 zmmword ptr [rip + head], zmm0
vzeroupper
ret
Clang 10 or ICC -O3 -mavx512f:
f(): # @f()
vmovdqa ymm0, ymmword ptr [rip + head]
vmovdqa ymm1, ymmword ptr [rip + head+32]
vmovdqa ymm2, ymmword ptr [rip + .LCPI0_0] # ymm2 =
[64,64,64,64,64,64,64,64,64,64,64,64,64,64,64,64]
vpsubusw ymm1, ymm1, ymm2
vpsubusw ymm0, ymm0, ymm2
vmovdqa ymmword ptr [rip + head], ymm0
vmovdqa ymmword ptr [rip + head+32], ymm1
vzeroupper
ret</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>