|
[Sponsors] |
February 27, 2007, 12:05 |
parallell computation with subroutine
|
#1 |
Guest
Posts: n/a
|
Hi guys,
I am using star-cd version4.0. I tried to run a case with newxyz and posdat subroutine, which is OK for one cup. But after I run the case with two cpus for a few seconds, I got the following error information. Anyone knows what is the problem? PNP: Initialized [2007-02-27-16:01:40] Automatic Sequential Automatic Parallel analyzer. PNP: Allocated "mace-r-ruixu,2" resource (2 processes). PNP: Assigned "mace-r-ruixu,2" to STAR solver (2 processes). PNP: Activated "ssh" command for starting tasks on remote nodes. PNP: Loading STAR double precision solver dynamic shared object plug-ins. PNP: Copyright Run Time Portions of this program include material copyrighted (C) by Absoft Corporation 1988-2006 PNP: Loading LAM MPI 7.0 dynamic shared object plug-ins. PNP: Loading STAR user coding dynamic shared object plug-in. PNP: Decomposing STAR mesh using "metis" method for 2 domains. PNP: Copying STAR master input file "run_up_WG1_114_25.ccm" to master domain. PNP: Copying STAR shared input file "run_up_WG1_114_25.prob" to domains. PNP: Copying STAR shared input file "beach_bottum_cell.dat" to domains. PNP: Copying STAR shared input file "beach_toe_cell.dat" to domains. PNP: Copying STAR shared input file "bottum_cell_y.dat" to domains. PNP: Copying STAR shared input file "focal_point_cell.dat" to domains. PNP: Copying STAR shared input file "focal_point_coordinates.dat" to domains. PNP: Starting TRACKER task on "mace-r-ruixu" for monitoring host and process failures. PNP: Spawning STAR processes on single node. PNP: Saving STAR merged output file "run_up_WG1_114_25.run" from domains. PNP: Saving STAR merged output file "run_up_WG1_114_25.info" from domains. PNP: Saving STAR master output file "run_up_WG1_114_25.ccm" from master domain. PNP: Saving STAR master output file "run_up_WG1_114_25.rsi" from master domain. PNP: Shutdown [2007-02-27-16:02:29] Execution aborted by request (SIGABRT) after 49 seconds (TOTAL ELAPSED TIME) Many thanks in advance! |
|
February 27, 2007, 12:45 |
Re: parallell computation with subroutine
|
#2 |
Guest
Posts: n/a
|
Have you ever run successfully in parallel without user subroutines? If not then you don't have things configured correctly. If you can run in parallel without user subroutines then the odds are that you are trying to do something in your subroutine that is illegal in parallel. For example, you can't loop over all the cells in the model in parallel - only all the cells in that processor.
|
|
February 27, 2007, 12:58 |
Re: parallell computation with subroutine
|
#3 |
Guest
Posts: n/a
|
I think I looped over some cells in the model.
I think that is the problem. But how can I find the cells which I need in my subroutine? Many thanks! |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Parallell cases | fabianpk | OpenFOAM Running, Solving & CFD | 0 | June 21, 2006 04:40 |
Computation time | Maciej | FLUENT | 3 | November 5, 2005 07:16 |
MPI and parallel computation | Wang | Main CFD Forum | 4 | April 3, 2005 07:40 |
computation bug? | Scott | Siemens | 3 | July 21, 2004 10:39 |
MPI and parallel computation | Wang | Main CFD Forum | 7 | April 15, 2004 12:25 |