hi,
I am still having the error that the cuda version of sander cannot find
the input file for REMD. Amber passes all the tests except for the REMD ones.
I am asking or 1 node, 2 tasks per node and 2 GPUS per node, what am I doing wrong?
thanks, Marcela
#SBATCH -N 1
#SBATCH --ntasks-per-node 2
#SBATCH -p GPU
#SBATCH --gres=gpu:p100:2
#
export DO_PARALLEL="mpirun -np 2"
module load cuda
export CUDA_HOME=/opt/packages/cuda/9.2
cd $SLURM_SUBMIT_DIR
make test >& maketestcudampi2.log
> Two replica GB REMD test.
>
> Running multipmemd version of pmemd Amber18
> Total processors = 2
> Number of groups = 2
>
>
> Unit 5 Error on OPEN: rem.in.001
>
> Unit 5 Error on OPEN: rem.in.001
> Abort(1) on node 1 (rank 1 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1
> ./Run.rem.sh: Program error
> make[5]: [rem_2rep_gb] Error 1 (ignored)
> --------------------------------------------------------------------------------
> Two replica PME REMD test.
>
> Running multipmemd version of pmemd Amber18
> Total processors = 2
> Number of groups = 2
>
>
> Unit 5 Error on OPEN: rem.in.001
>
> Unit 5 Error on OPEN: rem.in.001
> Abort(1) on node 1 (rank 1 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1
> ./Run.rem.sh: Program error
Marcela Madrid, Ph.D.
Senior Computational Scientist
Pittsburgh Supercomputing Center
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jul 01 2019 - 17:30:02 PDT
