I just meant without resorting to CPUID, because it is so unportable (to consider implementing directly into our library).
I've attached the output of x86info.
cpuinfo changed between RHEL 4.x and 5.x (easier to interpret in the latter w.r.t. SMT):
Is it possible to get at your CPUID backend now?
From: hwloc-users-bounces_at_[hidden] [mailto:hwloc-users-bounces_at_[hidden]] On Behalf Of Samuel Thibault
Sent: Thursday, January 28, 2010 10:24 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 09:30:06 -0700, a écrit :
> I've inlined /proc/cpuinfo at the end of this message.
Mmm, the different processors are really claimed to have different physical IDs.
Could you run x86info to easily get the APIC ID?
> Yes, the cache path you indicated doesn't exist!
> Is there any reliable alternate way to retrieve cache information
> (without CPUID) for those older kernels
Mmm, what do you mean by "without CPUID"?
I have implemented a cpuid-based backend for the systems that do not
provide a topology interface. I guess I may start leaning toward the
idea of being able to choose between plugins at runtime, since here it
could be useful.
hwloc-users mailing list