<html>
<head>
<base href="https://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 --- - KMP_TOPOLOGY_METHOD=cpuinfo not always works"
href="https://llvm.org/bugs/show_bug.cgi?id=27965">27965</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>KMP_TOPOLOGY_METHOD=cpuinfo not always works
</td>
</tr>
<tr>
<th>Product</th>
<td>OpenMP
</td>
</tr>
<tr>
<th>Version</th>
<td>unspecified
</td>
</tr>
<tr>
<th>Hardware</th>
<td>Other
</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>Runtime Library
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>pawel.osmialowski@foss.arm.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>KMP_TOPOLOGY_METHOD=cpuinfo implementation relies on presence of 'physical id'
field for each computation unit listed in /proc/cpuinfo file. Unfortunately,
this field seems to be x86-specific and is not present on many other
architectures. As a result, OpenMP program started with KMP_TOPOLOGY_METHOD
shell variable set to 'cpuinfo' on non-x86 ends like this:
OMP: Error #129: /proc/cpuinfo, line 8: cpuinfo file missing physical id field
- exiting.
Unfortunately, there's no solution portable across architectures for this. As
this occurs usually on embedded systems with single manycore physical CPU on
board, I'd suggest that in case of absence of 'physical id' field, all
computation units should be pinned to single core (i.e. core 0).</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>