Running DualSPHysics_v5.0 on HPC cluster
I could run DualSPHysics_v5.0 on boreas cluster.
I wrote a very simple shell script to submit the job to boreas.
It is possible to request the number of nods and cores and memory per core in the shell script.
The problem is when I increased the number of cores I would expect to have less run time- technically because I allocated more resources to the job- but surprisingly increasing number of cores has no effect on run time.
I would be appreciated if anyone could share any experience in that regard.
Best of luck to everybody