|Xeon E5-2680 v2 (10c)||6.3|
N=64017, rho=0.382, r_cut=3.0
|Xeon E5-2680 v2 (10c)||6.1|
N=64000, rho=0.382, r_cut=3.0
Performance numbers on this page are reported in millions of particle time steps per second (N*TPS). All benchmarks are executed in single precision. On Tesla GPUs, ECC is enabled and boost clocks are set to the maximum. Where noted, (Nc) CPU benchmarks use MPI to scale to all cores in a socket and (serial) executions run only on a single core.
NVIDIA provided the Tesla K40s and GTX 980 used in these benchmarks.