Weak scaling performance, Collide benchmark, Broadwell, 1M particles/node
Performance in millions of particle-timesteps / second / node

Nodes CPU (mpi,hyper) Kokkos/OMP (mpi,hyper,thread) Kokkos/serial (mpi,hyper)
1 321.5 (72,2) 242.9 (18,2,4) 271.6 (72,2)
2 286.3 (72,2) 245.6 (18,2,4) 253.5 (72,2)
4 305.6 (72,2) 221.1 (18,2,4) 243.8 (72,2)
8 297.4 (72,2) 202.3 (18,2,4) 266.4 (72,2)
16 296.6 (72,2) 187.9 (18,2,4) 254.9 (72,2)
32 265.5 (36,1) 179.8 (18,2,4) 227.5 (36,1)
64 260.3 (72,2) 175.1 (18,2,4) 181.3 (72,2)

Run commands and logfile links for column CPU

1 mpiexec -np 72 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 64 -v y 40 -v z 40 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=1.mpi=72.hyper=2
2 mpiexec -np 144 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 64 -v y 40 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=2.mpi=72.hyper=2
4 mpiexec -np 288 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 64 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=4.mpi=72.hyper=2
8 mpiexec -np 576 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 128 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=8.mpi=72.hyper=2
16 mpiexec -np 1152 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 128 -v y 80 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=16.mpi=72.hyper=2
32 mpiexec -np 1152 -npernode 36 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 128 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=32.mpi=36.hyper=1
64 mpiexec -np 4608 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_cpu -v x 256 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=cpu.kind=weak.size=1M.node=64.mpi=72.hyper=2

Run commands and logfile links for column Kokkos/OMP

1 mpiexec -np 18 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 40 -v z 40 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=1.mpi=18.thread=4.hyper=2
2 mpiexec -np 36 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 40 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=2.mpi=18.thread=4.hyper=2
4 mpiexec -np 72 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=4.mpi=18.thread=4.hyper=2
8 mpiexec -np 144 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=8.mpi=18.thread=4.hyper=2
16 mpiexec -np 288 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 80 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=16.mpi=18.thread=4.hyper=2
32 mpiexec -np 576 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=32.mpi=18.thread=4.hyper=2
64 mpiexec -np 1152 -npernode 18 --oversubscribe --bind-to socket ./spa_serrano_kokkos_omp -sf kk -k on t 4 -pk kokkos reduction parallel/reduce comm classic -v x 256 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_omp.kind=weak.size=1M.node=64.mpi=18.thread=4.hyper=2

Run commands and logfile links for column Kokkos/serial

1 mpiexec -np 72 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 40 -v z 40 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=1.mpi=72.hyper=2
2 mpiexec -np 144 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 40 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=2.mpi=72.hyper=2
4 mpiexec -np 288 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 64 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=4.mpi=72.hyper=2
8 mpiexec -np 576 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 80 -v z 80 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=8.mpi=72.hyper=2
16 mpiexec -np 1152 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 80 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=16.mpi=72.hyper=2
32 mpiexec -np 1152 -npernode 36 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 128 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=32.mpi=36.hyper=1
64 mpiexec -np 4608 -npernode 72 --oversubscribe --bind-to core ./spa_serrano_kokkos_serial -sf kk -k on -pk kokkos reduction parallel/reduce comm classic -v x 256 -v y 160 -v z 160 -v t 100 -in in.collide.steps -log log.sparta.date=23Dec17.model=collide.machine=serrano.pkg=kokkos_serial.kind=weak.size=1M.node=64.mpi=72.hyper=2