<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 - ThinLTO doesn't use all virtual cores as advertised"
href="https://bugs.llvm.org/show_bug.cgi?id=38610">38610</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>ThinLTO doesn't use all virtual cores as advertised
</td>
</tr>
<tr>
<th>Product</th>
<td>new-bugs
</td>
</tr>
<tr>
<th>Version</th>
<td>unspecified
</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>new bugs
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>mh+llvm@glandium.org
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr></table>
<p>
<div>
<pre>The ThinLTO documentation (<a href="https://clang.llvm.org/docs/ThinLTO.html">https://clang.llvm.org/docs/ThinLTO.html</a>) says:
By default, the ThinLTO link step will launch up to
std::thread::hardware_concurrency number of threads in parallel. For machines
with hyper-threading, this is the total number of virtual cores.
That appears not to be true. On a machine with 8 cores and 16 hyper-threads,
std::thread::hardware_concurrency returns 16, but the linkage (with lld) only
uses 8 threads. I have to manually pass --thinlto-jobs=16 for 16 threads to be
used. This does make a difference, where the linkage with 16 threads takes 75%
of the time the one with 8 threads takes.
Tangentially, GCC has (opt-in) Make jobserver support, which is even better.</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>