[CIG-MC] CitcomCU with 100 procs

[Shijie Zhong]

Those arrays may not be needed. We will probably just remove them from the code soon. 

Shijie

Shijie Zhong
Department of Physics
University of Colorado at Boulder
Boulder, CO 80309
Tel: 303-735-5095; Fax: 303-492-7935
Web: http://anquetil.colorado.edu/szhong


---- Original message ----
>Date: Tue, 18 Mar 2008 12:28:36 -0700
>From: Eh Tan <tan2 at geodynamics.org>  
>Subject: Re: [CIG-MC] CitcomCU with 100 procs  
>To: Tobias Höink <tobias.hoeink at rice.edu>
>Cc: cig-mc at geodynamics.org
>
>Hi Tobias,
>
>(cc to cig-mc list)
>
>I recently got the same error. After digging into the code, I found that 
>there is a hard-coded limit on the number of processrs. The limit is 
>currently 100. To increase the limt to 200, at line 470, global_def.h, 
>change these lines:
>
>        int mst[100][100][3];
>        int mst1[100][100];
>        int mst2[100][100];
>        int mst3[100][100];
>
>to
>        int mst[200][200][3];
>        int mst1[200][200];
>        int mst2[200][200];
>        int mst3[200][200];
>
>
>Eh
>
>
>Tobias Höink wrote:
>
>> Hi Eh,
>>
>> I have tried running CitcomCU on a large number of processors. It  
>> works well with 64 procs, but not with 100 procs.
>> Based on the output (program stopped after 'ok15') I assume that  
>> something in mass_matrix() is causing problems.
>> Would you know where the problem is? Please find output below and  
>> input file attached.
>>
>> Thanks,
>> Tobias
>>
>> ok1
>> ok2
>> ok3
>> Problem has 321 x 49 x 321 nodes
>> ok4
>> ok5
>> ok6
>> ok6a
>> ok9
>> ok10
>> ok11
>> ok12
>> ok7a
>> ok13
>> ok14
>> ok15
>> =>> PBS: job killed: walltime 1826 exceeded limit 1800
>> Cleaning up before exit
>> [...]
>>
>
>_______________________________________________
>CIG-MC mailing list
>CIG-MC at geodynamics.org
>http://geodynamics.org/cgi-bin/mailman/listinfo/cig-mc