Sorry, I'm probably lumping more than I should under the name "cpuid".
Are you using an algorithm like the one recommended by Intel for your x86 backend? http://software.intel.com/en-us/articles/optimal-performance-on-multithreaded-software-with-intel-tools/ Thanks :) -----Original Message----- From: hwloc-users-boun...@open-mpi.org [mailto:hwloc-users-boun...@open-mpi.org] On Behalf Of Samuel Thibault Sent: Thursday, January 28, 2010 11:46 AM To: Hardware locality user list Subject: Re: [hwloc-users] No caches or hiearchy on RHEL 4.7 or 4.8 Dan Eaton, le Thu 28 Jan 2010 11:36:49 -0700, a écrit : > Yes, the cpuid version works marvelously. The only thing it misses is > memory/NUMA node (is it using libnuma?). Cpuids do not have any relation with NUMA nodes. Thanks for the sample, it is a hint that maybe we'll have to permit users & applications to be quite precise about which backend should detect what (e.g. here, blacklist linux' core topology information, but use its numa information, and use the core and cache topology information from the x86 backend). Samuel _______________________________________________ hwloc-users mailing list hwloc-us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users