. 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

Table 2.

Achieved n-fold speed-up from multi-GPU execution of benchmark systems.

Device	Spacing	GPUs	E. coli	Dividing E. coli	Yeast	Red Blood Cell
GTX680	16 nm	2	1.83	1.91	1.98	1.99
GTX680	16 nm	4	2.84	3.30	3.78	3.91
M2070	16 nm	2	1.74	1.85	1.98	1.98
		4	2.04	2.64	3.79	3.91
		8	1.56	2.37	4.04	6.79
GTX680	8 nm	2	1.96	1.96	1.99	-
GTX680	8 nm	4	3.61	3.81	3.88	-
M2070	8 nm	2	1.92	1.96	1.99	-
		4	3.27	3.59	3.91	-
		8	3.08	3.77	6.36	-

Large systems exhibit nearly linear scaling, with smaller systems performing well on lesser numbers of GPUs. For the red blood cell test, 8nm speed-up data is not available because it to large be run on one GPU for comparison.