Open MPI logo

Hardware Locality Users' Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Hardware Locality Users mailing list

Subject: Re: [hwloc-users] Strange binding issue on 40 core nodes and cgroups
From: Brock Palen (brockp_at_[hidden])
Date: 2012-11-05 16:57:23


Ok more information (had to build newer hwloc) My job today only 2 processes are running at half speed and they indeed are sharing the same core:

[root_at_nyx7000 ~]# for x in `cat /tmp/pids `; do echo -n "$x "; hwloc-bind --get-last-cpu-location --pid $x; done | sort -k 2
1164 0x00000001,0x0
1158 0x00000010,0x0
1165 0x00000010,0x0
1167 0x00000020
1157 0x00000200
1159 0x00000400
1160 0x00002000
1163 0x00004000
1166 0x00020000
1161 0x00040000
1168 0x00200000
1162 0x00400000

 1157 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 9:49.55 stream
 1159 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 8:10.91 stream
 1161 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 9:49.55 stream
 1162 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 9:49.54 stream
 1163 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 9:49.55 stream
 1164 brockp 20 0 1885m 1.8g 456 R 99.6 0.2 9:49.53 stream
 1160 brockp 20 0 1885m 1.8g 456 R 97.7 0.2 9:49.54 stream
 1166 brockp 20 0 1885m 1.8g 456 R 97.7 0.2 9:49.53 stream
 1167 brockp 20 0 1885m 1.8g 456 R 97.7 0.2 9:49.46 stream
 1168 brockp 20 0 1885m 1.8g 456 R 97.7 0.2 8:10.86 stream
 1158 brockp 20 0 1885m 1.8g 456 R 48.9 0.2 4:54.78 stream
 1165 brockp 20 0 1885m 1.8g 456 R 48.9 0.2 4:54.76 stream

This is very strange. Is there a way to ask hwloc to show me all processes that are using a given cpu?

Brock Palen
www.umich.edu/~brockp
CAEN Advanced Computing
brockp_at_[hidden]
(734)936-1985

On Nov 2, 2012, at 4:30 PM, Brice Goglin wrote:

> Le 02/11/2012 21:22, Brice Goglin a écrit :
>> hwloc-bind --get-last-cpu-location --pid <pid> should give the same
>> info but it seems broken on my machine right now, going to debug.
>
> Actually, that works fine once you try it on a non-multithreaded program
> that uses all cores :)
>
> So you can use top or hwloc-bind --get-last-cpu-location --pid <pid> to
> find out where each process runs.
>
> Brice
>
> _______________________________________________
> hwloc-users mailing list
> hwloc-users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users