<html>
<head>
<base href="https://bugs.freedesktop.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - Register spilling is slow"
href="https://bugs.freedesktop.org/show_bug.cgi?id=70613#c6">Comment # 6</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW --- - Register spilling is slow"
href="https://bugs.freedesktop.org/show_bug.cgi?id=70613">bug 70613</a>
from <span class="vcard"><a class="email" href="mailto:kevin.rogovin@intel.com" title="Kevin Rogovin <kevin.rogovin@intel.com>"> <span class="fn">Kevin Rogovin</span></a>
</span></b>
<pre>The register spilling is a product of the fact that the noise functions are
inlined without barriers of the inline. If a function was tagged with how many
registers it required then those registers could be reused at its next
invocation. Instead, in an attempt to optimize via inlining, that barrier is
lost and register allocation and freeing is far messier to track. Indeed, that
it compiles quickly with one invocation provides evidence that register use
tracking is the issue... that and over inlining of functions.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the QA Contact for the bug.</li>
</ul>
</body>
</html>