[cfMesh] cfMesh - OpenFOAM v1812 not working in parallel

codygo · March 8, 2019, 14:21

My case is approximately 60M cells, and meshes fairly quickly in cfMesh if I don't attempt "-parallel" (preceded by preparePar, using proper mpi call)

Code:

[qtzlctl:32171] *** An error occurred in MPI_Recv
[qtzlctl:32171] *** reported by process [3578789889,16]
[qtzlctl:32171] *** on communicator MPI_COMM_WORLD
[qtzlctl:32171] *** MPI_ERR_TRUNCATE: message truncated
[qtzlctl:32171] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[qtzlctl:32171] ***    and potentially your MPI job)

I've tried varying MPI_BUFFER_SIZE, and though it does seem to affect where it crashes, I haven't been able to find a stable value.

I can get a little closer if I attempt to run preparePar after creating the mesh, but this is painfully slow.

I'd like to be able to get a successful parallel mesh with cfMesh, and continue adding layers using snappyHexMesh.

olivierG · May 3, 2019, 07:22

hello,
CfMesh work in parallel without MPI on a multi core single machine. So if your PC have 8 core, set OMP_NUM_THREADS=8 and run CfMesh.
Usually, you just need this.

The only case you would use MPI parallel meshing, is in case you run CfMesh on 2 or more PC (Cluster). In that case, you should infor MPI chich machine and how many core in your cluster. But your error message does not permit to know what 's wrong in your case.

regards,
olivier

March 8, 2019, 14:21	cfMesh - OpenFOAM v1812 not working in parallel	#1
codygo New Member cody Join Date: Mar 2009 Posts: 9 Rep Power: 17	My case is approximately 60M cells, and meshes fairly quickly in cfMesh if I don't attempt "-parallel" (preceded by preparePar, using proper mpi call) Code: [qtzlctl:32171] * An error occurred in MPI_Recv [qtzlctl:32171] * reported by process [3578789889,16] [qtzlctl:32171] * on communicator MPI_COMM_WORLD [qtzlctl:32171] * MPI_ERR_TRUNCATE: message truncated [qtzlctl:32171] * MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort, [qtzlctl:32171] * and potentially your MPI job) I've tried varying MPI_BUFFER_SIZE, and though it does seem to affect where it crashes, I haven't been able to find a stable value. I can get a little closer if I attempt to run preparePar after creating the mesh, but this is painfully slow. I'd like to be able to get a successful parallel mesh with cfMesh, and continue adding layers using snappyHexMesh.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Error running openfoam in parallel	fede32	OpenFOAM Programming & Development	5	October 4, 2018 17:38
Openfoam parallel calculation performance study - Half performance on mpirun	Jurado	OpenFOAM Running, Solving & CFD	22	March 24, 2018 21:40
OpenFOAM v3.0.1 Training, London, Houston, Berlin, Jan-Mar 2016	cfd.direct	OpenFOAM Announcements from Other Sources	0	January 5, 2016 04:18
OpenFOAM Training, London, Chicago, Munich, Sep-Oct 2015	cfd.direct	OpenFOAM Announcements from Other Sources	2	August 31, 2015 14:36
simpleFoam parallel	AndrewMortimer	OpenFOAM Running, Solving & CFD	12	August 7, 2015 19:45

May 3, 2019, 07:22		#2
olivierG Senior Member Olivier Join Date: Jun 2009 Location: France, grenoble Posts: 272 Rep Power: 18	hello, CfMesh work in parallel without MPI on a multi core single machine. So if your PC have 8 core, set OMP_NUM_THREADS=8 and run CfMesh. Usually, you just need this. The only case you would use MPI parallel meshing, is in case you run CfMesh on 2 or more PC (Cluster). In that case, you should infor MPI chich machine and how many core in your cluster. But your error message does not permit to know what 's wrong in your case. regards, olivier