Welcome! Log In Create A New Profile



Posted by bohisbohis  
May 02, 2018 06:56AM
Dear all!

in Palabos I am struggling with saving data when running a parallel job on cluster with 100cores.
In the beginning of the simulation I am reading
1) the boolean mask form ASCII file
2)loadBinaryBlock(nsLattice, "checkpointTemp.dat");
loadBinaryBlock(adLattice, "checkpointVel.dat");


at the end of the simulation I have to save all data by doing:

pcout << "writing CheckpointTemp.dat ... " << endl;
saveBinaryBlock(nsLattice, "checkpointTemp.dat");
pcout << "writing CheckpointVel.dat ... " << endl;
saveBinaryBlock(adLattice, "checkpointVel.dat");

Interestingly, sometimes if I resubmit the same job, the task will be completed successfully, more often it however only generates an empty file "checkpointTemp.dat" and the job continues till the walltime is consumed.

I don't receive any error message, not even when Debug=true.

Anyone's idea would be greatly appreciated, any comments?

Sorry, you do not have permission to post/reply in this forum.