<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 --- - Regression due to increased register pressure after r172868 revision."
href="http://llvm.org/bugs/show_bug.cgi?id=17285">17285</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>Regression due to increased register pressure after r172868 revision.
</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>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>llc
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>ili.filippov@gmail.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>Created <span class=""><a href="attachment.cgi?id=11236" name="attach_11236" title="Reproducer">attachment 11236</a> <a href="attachment.cgi?id=11236&action=edit" title="Reproducer">[details]</a></span>
Reproducer
Revision r172868: "On Sandybridge loading unaligned 256bits using two XMM loads
(vmovups and vinsertf128) is faster than using a single vmovups instruction."
increases register pressure, which becomes critical in case of multiple loads.
This shows up on many ISPC tests with total test slow down up to 1.5x (due to
increased number of spill/fills).
An attached example is extracted from one of these tests and demonstrated the
problem.
I don't know how much faster loading through XMM on Sandybridge is, but in case
of multiple loads it's surely isn't.</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>