Graphical processing units are rapidlygaining maturity as powerful general parallelcomputing devices. The package cudaBayesreguses GPU-oriented procedures to improve theperformance of Bayesian computations. Thepaper motivates the need for devising highperformance computing strategies in the context of fMRI data analysis. Some features of thepackage for Bayesian analysis of brain fMRI dataare illustrated. Comparative computing performance ﬁgures between sequential and parallelimplementations are presented as well.
|Journal||The R Journal|
|Publication status||Published - 1 Jan 2010|