[mvapich-discuss] MV2_CPU_MAPPING doesn't work on all the nodes

Jonathan Perkins perkinjo at cse.ohio-state.edu
Mon Jun 25 11:32:47 EDT 2012


Hello, can you try using mpirun_rsh to see if you get different
results.  Your command should be...

mpirun_rsh -n 32 -hostfile mfile.1 MV2_CPU_MAPPING 0:1:2:3:6:7:8:9 ./namd2

On Fri, Jun 22, 2012 at 11:04 PM, Ju JiaJia <jujj603 at gmail.com> wrote:
> Hi all:
>
> I'm running a program with 32 processes on 4 nodes using 8 cores each . Each
> node has two 6-core Xeon processor and all nodes is free. Running command:
> mpiexec -n 32 -hostfile mfile.1 -env MV2_CPU_MAPPING 0:1:2:3:6:7:8:9 ./namd2
> apoa1.namd
>
> Here is mfile.1:
> [Sat Jun 23 10:41:38]user39 at m1:bin$ cat mfile.1
> c04b04:8
> c04b05:8
> c04b06:8
> c04b07:8
>
> After subbmitting the job, i logon to the nodes in mfile.1, i find not all
> nodes use the cores as i specified. C04b04 use  0:1:2:3:6:7:8:9, but the
> last three don't. I didn't change MV2_ENABLE_AFFINITY, so it's 1 by default.
> Why MV2_CPU_MAPPING won't work ?
>
> _______________________________________________
> mvapich-discuss mailing list
> mvapich-discuss at cse.ohio-state.edu
> http://mail.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>



-- 
Jonathan Perkins
http://www.cse.ohio-state.edu/~perkinjo



More information about the mvapich-discuss mailing list