#!/bin/bash

#PBS -l select=1:system=sunspot,place=scatter
#PBS -A LatticeQCD_aesp_CNDA
#PBS -l walltime=01:00:00
#PBS -N dwf
#PBS -k doe

HDIR=/home/paboyle/
module use /soft/testing/modulefiles/
module load intel-UMD23.05.25593.11/23.05.25593.11
module load tools/pti-gpu  
export LD_LIBRARY_PATH=$HDIR/tools/lib64:$LD_LIBRARY_PATH
export PATH=$HDIR/tools/bin:$PATH

export TZ='/usr/share/zoneinfo/US/Central'
export OMP_PROC_BIND=spread
export OMP_NUM_THREADS=3
unset OMP_PLACES

cd $PBS_O_WORKDIR

qsub jobscript.pbs

echo Jobid: $PBS_JOBID
echo Running on host `hostname`
echo Running on nodes `cat $PBS_NODEFILE`

echo NODES
cat $PBS_NODEFILE
NNODES=`wc -l < $PBS_NODEFILE`
NRANKS=12         # Number of MPI ranks per node
NDEPTH=4          # Number of hardware threads per rank, spacing between MPI ranks on a node
NTHREADS=$OMP_NUM_THREADS # Number of OMP threads per rank, given to OMP_NUM_THREADS

NTOTRANKS=$(( NNODES * NRANKS ))

echo "NUM_NODES=${NNODES}  TOTAL_RANKS=${NTOTRANKS}  RANKS_PER_NODE=${NRANKS}  THREADS_PER_RANK=${OMP_NUM_THREADS}"
echo "OMP_PROC_BIND=$OMP_PROC_BIND OMP_PLACES=$OMP_PLACES"

    
CMD="mpiexec -np ${NTOTRANKS} -ppn ${NRANKS} -d ${NDEPTH} --cpu-bind=depth -envall \
	     ./gpu_tile_compact.sh \
	./Benchmark_dwf_fp32 --mpi 1.1.2.6 --grid 16.32.64.192 --comms-overlap \
	--shm-mpi 0 --shm 2048 --device-mem 32000 --accelerator-threads 32"