<html>
<head>
<base href="http://llvm.org/bugs/" />
</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 CodeGen: Vector code for shift right could be better?"
href="http://llvm.org/bugs/show_bug.cgi?id=15627">15627</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>X86 CodeGen: Vector code for shift right could be better?
</td>
</tr>
<tr>
<th>Product</th>
<td>tools
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</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>llc
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>aschwaighofer@apple.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvmbugs@cs.uiuc.edu
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>#define TYPE unsigned int
#define OP >>
#define SIZE 1024
#define TYPE_ALIGN __attribute__((aligned(16)))
TYPE A1[SIZE] TYPE_ALIGN;
TYPE B1[SIZE] TYPE_ALIGN;
TYPE C1[SIZE] TYPE_ALIGN;
TYPE D1[SIZE] TYPE_ALIGN;
void kernel1(int x) {
for (int i = 0; i < SIZE; ++i) {
//A1[i] = B1[i] OP C1[i];
A1[i] = B1[i] OP x;
}
}
Couldn't we use psrlw instead of the scalarized shr.'s?
cat > testshr.ll
define <4 x i32> @test_shr(<4 x i32> %a, <4 x i32> %b) {
%1 = lshr <4 x i32> %a, %b
ret <4 x i32> %1
}
----
llc testshr.ll -o - -mcpu=core2 -mtriple=x86_64-apple-macosx10.8.0
## BB#0:
pshufd $3, %xmm0, %xmm2 ## xmm2 = xmm0[3,0,0,0]
movd %xmm2, %edx
pshufd $3, %xmm1, %xmm2 ## xmm2 = xmm1[3,0,0,0]
movd %xmm2, %ecx
## kill: CL<def> CL<kill> ECX<kill>
shrl %cl, %edx
movd %xmm0, %eax
movd %xmm1, %ecx
## kill: CL<def> CL<kill> ECX<kill>
shrl %cl, %eax
movd %edx, %xmm2
pshufd $1, %xmm0, %xmm3 ## xmm3 = xmm0[1,0,0,0]
movd %xmm3, %edx
pshufd $1, %xmm1, %xmm3 ## xmm3 = xmm1[1,0,0,0]
movd %xmm3, %ecx
## kill: CL<def> CL<kill> ECX<kill>
shrl %cl, %edx
movd %edx, %xmm3
punpckldq %xmm2, %xmm3 ## xmm3 =
xmm3[0],xmm2[0],xmm3[1],xmm2[1]
movd %eax, %xmm2
movhlps %xmm0, %xmm0 ## xmm0 = xmm0[1,1]
movd %xmm0, %eax
movhlps %xmm1, %xmm1 ## xmm1 = xmm1[1,1]
movd %xmm1, %ecx
## kill: CL<def> CL<kill> ECX<kill>
shrl %cl, %eax
movd %eax, %xmm0
punpckldq %xmm0, %xmm2 ## xmm2 =
xmm2[0],xmm0[0],xmm2[1],xmm0[1]
punpckldq %xmm3, %xmm2 ## xmm2 =
xmm2[0],xmm3[0],xmm2[1],xmm3[1]
movdqa %xmm2, %xmm0
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>