Item response theory (IRT) is a popular approach used for addressing statistical problems in psychometrics as well as in other\r\nfields. The fully Bayesian approach for estimating IRT models is computationally expensive. This limits the use of the procedure\r\nin real applications. In an effort to reduce the execution time, a previous study shows that high performance computing provides\r\na solution by achieving a considerable speedup via the use of multiple processors. Given the high data dependencies in a single\r\nMarkov chain for IRT models, it is not possible to avoid communication overhead among processors. This study is to reduce\r\ncommunication overhead via the use of a row-wise decomposition scheme. The results suggest that the proposed approach\r\nincreased the speedup and the efficiency for each implementation while minimizing the cost and the total overhead. This further\r\nsheds light on developing high performance Gibbs samplers for more complicated IRT models.
Loading....