Skip to main content
. Author manuscript; available in PMC: 2014 May 30.
Published in final edited form as: Parallel Comput. 2014 Apr 2;40(5-6):86–99. doi: 10.1016/j.parco.2014.03.009

Figure 5.

Figure 5

Time required for simulating one hour of the performance test systems using multiple GPUs. Runtime grows with system volume in a nearly linear fashion. The simulations were run using four GTX680 GPUs and also with eight M2070 GPUs. The test systems are spatially homogeneous volumes simulating four reactions with four species (A ⇌ B, B + C ⇌ D) at constant density, with initial particle counts given in Table 1. Performance exhibits linear scaling as simulation volume increases on two and four GPUs, whereas using eight GPUs for the two smallest systems shows no benefit due to insufficient work per GPU.