Hybrid Parallel Computation of Integration in GRACE
Abstract
With an integrated software package {\tt GRACE}, it is possible to generate Feynman diagrams, calculate the total cross section and generate physics events automatically. We outline the hybrid method of parallel computation of the multi-dimensional integration of {\tt GRACE}. We used {\tt MPI} (Message Passing Interface) as the parallel library and, to improve the performance we embedded the mechanism of the dynamic load balancing. The reduction rate of the practical execution time was studied.