Activity for Prashant Gaikwad
-
Prashant Gaikwad posted a comment on discussion Elk Users
Hi. I am using 2 nodes with each node having 24 cores and batch file for SLURM. I have tried and played with settings given by you but it seems to be not working. Whenever I set --ntasks-per-node=1 then number of CPUs getting selected automatically= 1 per node with around 1200% utilization (by top command). So when I tried --ntaks-per-node=24 and OMP threads to be 24 then I can see that 24 cores (with around 100% utilization) on each node is engaged. However now total number of threads are 48x24=1152....
-
Hello Pavlov Nikita, thank you so very much. I will try these suggestions and get back to you. Big time help. Thanks.
-
In parallel run; multiple copies of ELK code are running individually. Hello, I recently started using ELK. Installation is done on cluster using mpiifort and I am trying to use slurm to submit the job. Since VASP and other softwares are running correctly, hence I believe all the mpi libraries including mkl are in order. When I am running program on single node using 12 cores for one of the example file for structural optimization of Ge, mpirun -np 12 elk <elk.in> ELK.OUT &</elk.in> I can see following...
-
In parallel run; multiple copies of ELK code are running individually. Hello, I recently started using ELK. Installation is done on cluster using mpiifort and I am trying to use slurm to submit the job. Since VASP and other softwares are running correctly, hence I believe all the mpi libraries including mkl are in order. When I am running program on single node using 12 cores for one of the example file for structural optimization of Ge, mpirun -np 12 elk <elk.in> ELK.OUT &</elk.in> I can see following...
-
In parallel run; multiple copies of ELK code are running individually. Hello, I recently started using ELK. Installation is done on cluster using mpiifort and I am trying to use slurm to submit the job. Since VASP and other softwares are running correctly, hence I believe all the mpi libraries including mkl are in order. When I am running program on single node using 12 cores for one of the example file for structural optimization of Ge, mpirun -np 12 elk <elk.in> ELK.OUT &</elk.in> I can see following...
1
