Skip to main content
. 2014 Oct 30;2014:368149. doi: 10.1155/2014/368149

Figure 5.

Figure 5

Execution time and speedup for implementing MPI and CUDA parallel programs of Gibbs sampling for tests with k = 100 items.