<html>
<head>
<base href="https://bugs.freedesktop.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - GLSL compilation can be very slow"
href="https://bugs.freedesktop.org/show_bug.cgi?id=93681#c3">Comment # 3</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - GLSL compilation can be very slow"
href="https://bugs.freedesktop.org/show_bug.cgi?id=93681">bug 93681</a>
from <span class="vcard"><a class="email" href="mailto:mattst88@gmail.com" title="Matt Turner <mattst88@gmail.com>"> <span class="fn">Matt Turner</span></a>
</span></b>
<pre>Created <span class=""><a href="attachment.cgi?id=120995" name="attach_120995" title="new.shader_test">attachment 120995</a> <a href="attachment.cgi?id=120995&action=edit" title="new.shader_test">[details]</a></span>
new.shader_test
(In reply to Matt Turner from <a href="show_bug.cgi?id=93681#c2">comment #2</a>)
<span class="quote">> I read the shader a bit more and have decided that all indirect accesses
> could be statically determined to occur in limited ranges of the temps[]
> array.</span >
With that knowledge in hand, I've modified slow.shader_test to limit indirect
accesses to small arrays -- by making
vec4 temps_9_14[6];
vec4 temps_15_20[6];
vec4 temps_21_25[5];
vec4 temps_26_30[5];
and modifying the accesses accordingly. That shader (attached with this
comment) compiles in the blink of an eye and is
712 instructions. 8 loops. 1146286 cycles. 0:0 spills:fills. 115 basic blocks.</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>