CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Large-scale simulation problem (

Purushothama November 7, 2010 21:12

Large-scale simulation problem
1 Attachment(s)
Dear colleagues,

I am using starcd 3.26 with ufiles for fuel cell simulation. I have observed some problem while simulating large-scale geometry having around 9 million cells. I recieved sigpipe error after a iteration and here below i am attaching the screenshot. I am using parallel compuatation with 48 GB main server and 8 cores (each has 4 processors) with 4GB memory each and I decompose the geometry manually in Y direction using set files. The command i issued is

star -dp -mpi=mpich -decompmeth=s node1,4 node2,4......

The dynamic memory allocated is about 900 MB.

When i run the simulation in one processor using ''star -dp'' command., it runs fine, but very very slow. I face problem only during parallel run. I have tried parallel run for other simple geometries with 1-1.5 millions of cells, everthing is fine with these simulations. Only when the cell number exceeds 2 million, the simulation diverges.

I would appreciate if anyone can help me to solve this problem.

Thanks in advance.

All times are GMT -4. The time now is 11:52.