Merge branch 'add_hybrid' into dev

test_scripts/occigen/hybrid/hsw24_or_bdw28_nodes.sh

+#!/bin/bash
+#SBATCH -J or_hybrid
+#SBATCH -N 3
+#SBATCH --threads-per-core=1
+#SBATCH --cpus-per-task=4
+#SBATCH --time=0:30:00
+#SBATCH --exclusive
+#SBATCH --mem=50GB
+#SBATCH --output or_hybrid.output.slurm.%J
+
+#OR
+#SBATCH -C HSW24|BDW28
+set -e
+
+export I_MPI_DOMAIN=auto
+export I_MPI_PIN_RESPECT_CPUSET=0
+export I_MPI_DEBUG=4
+
+#Make sure that OMP_NUM_THREADS = cpus-per-task * KMP_HW_SUBSET
+export KMP_HW_SUBSET=1T
+export OMP_NUM_THREADS=4
+export KMP_AFFINITY=verbose,compact,1,0,granularity=fine
+
+module load intel intelmpi
+
+ulimit -s unlimited
+
+export NCPUS=$((($SLURM_CPUS_ON_NODE/$OMP_NUM_THREADS)*$SLURM_NNODES))
+
+srun -n $NCPUS ../../../bin/hello_hybrid
+

test_scripts/occigen/hybrid/hsw24_or_bdw28_ntasks.sh

+#!/bin/bash
+#SBATCH -J or_hybrid
+#SBATCH --ntasks=10
+#SBATCH --threads-per-core=1
+#SBATCH --cpus-per-task=4
+#SBATCH --time=0:30:00
+#SBATCH --exclusive
+#SBATCH --mem=50GB
+#SBATCH --output or_hybrid.output.slurm.%J
+
+#OR
+#SBATCH -C BDW28|HSW24
+set -e
+
+export I_MPI_DOMAIN=auto
+export I_MPI_PIN_RESPECT_CPUSET=0
+export I_MPI_DEBUG=4
+
+#Make sure that OMP_NUM_THREADS = cpus-per-task * KMP_HW_SUBSET
+export KMP_HW_SUBSET=1T
+export OMP_NUM_THREADS=4
+export KMP_AFFINITY=verbose,compact,1,0,granularity=fine
+
+module load intel intelmpi
+
+ulimit -s unlimited
+
+srun ../../../bin/hello_hybrid
+

test_scripts/occigen/hybrid/hsw24_xor_bdw28_ntasks.sh

@@ -18,7 +18,6 @@ export I_MPI_DEBUG=4
 
 #Make sure that OMP_NUM_THREADS = cpus-per-task * KMP_HW_SUBSET
 export KMP_HW_SUBSET=1T
-#export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
 export OMP_NUM_THREADS=4
 export KMP_AFFINITY=verbose,compact,1,0,granularity=fine