560987.fig.0010
Figure 10: Speedup of the execution time per iteration of the linear stencil versus number of GPU cards. Done on Cluster 2 and resource 3.