Evaluation of Parallel Computing on MPI Version PHITS Code

The Message Passing Interface (MPI) technique is an old solution and improvement on the Monte Carlo N-Particle Transport (MCNP) method’s enormous computational time, which has not been evaluated based PHITS code—a recently developed simulation code. We conducted simulations Varian Clinac iX 6MV phase space data from IAEA. Venselaar et al.’s method criteria were used to validate simulation. PC cluster also tested in terms of processor count bch, stands for unit calculation per operation. MPI version code’s speedup factor K-factor, represent serial portion cluster, both evaluated. All calculated met except δ2, high dose, gradient beam profile set. It was very clear that clusters with better than simple nodes up 70.6%. Additionally, shows a tendency follow Amdahl’s law. At same K-factor saturated by certain measure. study concludes arguing limitations come its composition. If we consider how improvements specifications affect this system could be more effective.