Purpose of the -parallel parameter
Hello,
I am running a parallel job and I forgot the -parallel parameter, so my command to start the simulation looked like this mpirun -np 4 twoPhaseEulerFoam The simulation runs and the system monitor tells me that there are in fact 4 processes running. The only difference to a normal parallel run is, that in the processor* folders there is only a 0 and a constant folder. The computed time steps are stored in the case directory. Can anyone describe what exactly happens if I omit the -parallel parameter? I guess storing the field-information every time step requires some time, but running reconstructPar after the computation also does. |
without -parallel the processors will run the same job (in serial)
try running 'mpirun -np 4 ls' it will run the same ls-command 4 times, on each processor. |
Thanks
your example is very enlightening |
All times are GMT -4. The time now is 02:20. |