forked from portelli/lattice-benchmarks
working scripts for Tursa
This commit is contained in:
14
Grid/systems/tursa/files/gpu-mpi-wrapper.sh
Normal file
14
Grid/systems/tursa/files/gpu-mpi-wrapper.sh
Normal file
@ -0,0 +1,14 @@
|
||||
#!/usr/bin/env bash
|
||||
|
||||
lrank=$OMPI_COMM_WORLD_LOCAL_RANK
|
||||
numa1=$(( 2 * lrank))
|
||||
numa2=$(( 2 * lrank + 1 ))
|
||||
netdev=mlx5_${lrank}:1
|
||||
|
||||
export CUDA_VISIBLE_DEVICES=$OMPI_COMM_WORLD_LOCAL_RANK
|
||||
export UCX_NET_DEVICES=${netdev}
|
||||
BINDING="--interleave=$numa1,$numa2"
|
||||
|
||||
echo "$(hostname) - $lrank device=$CUDA_VISIBLE_DEVICES binding=$BINDING"
|
||||
|
||||
numactl ${BINDING} "$@"
|
Reference in New Issue
Block a user