Am 07.03.2011 um 15:26 schrieb Esztermann, Ansgar:

> On Mar 7, 2011, at 15:09 , Reuti wrote:
> 
>> Am 07.03.2011 um 14:42 schrieb Esztermann, Ansgar:
>> 
>>> Hi List,
>>> 
>>> is anyone using the core binding feature on AMD Magny-Cours? If so, how do 
>>> you do it? For me, loadcheck mis-reads the topology as 
>>> (SCTTCTTCTTCTTCTTCTT)*4 rather than (SCCCCCCCCCCCC)*4. As far as I can 
>>> tell, there are two problems:
>>> - the kernel (2.6.18-194.26.1.el5) only exposes core_id and 
>>> physical_package_id for each core, with core_id running from 0 to 5 and 
>>> physical_package_id from 1 to 4. However, there are two dice (nodes) per 
>>> socket, but this is not shown in the device tree.
>> 
>> Confirmed. But I wonder, whether it's SGE's fault. Why does the kernel 
>> ouptut the same core id, although it's not the same core?
> 
> Well, core IDs are unique only within the same socket ID (for older CPUs, say 
> Harpertown),

Yep. But it's working fine this way with former CPUs.


> so I would assume the same holds for node IDs -- it's just that node IDs 
> aren't displayed for Magny-Cours.

It's the physical ID which is changing for each socket. Magny-Cours just output 
doubles for "physical id" and "core id" entries. But the "apicid" and "initial 
apicid" are changing. But I don't know what's happening for Intel CPUs for 
these entries.

Isn't it an issue of any kernel list?

http://us.generation-nt.com/answer/patch-x86-wrong-proc-cpuinfo-core-id-amd-fam-f-model-9-help-199827011.html

Looks like others found this too...

-- Reuti


> A.
> 
> -- 
> Ansgar Esztermann
> DV-Systemadministration
> Max-Planck-Institut für biophysikalische Chemie, Abteilung 105
> 
> 


_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users

Reply via email to