Parallelization of KT3D Application of GSLIB Library Using Coprocessor Intel Xeon Phi with OpenMP
Alfonso Ferreira, F.F.A. Ponte, L.E.S. Varella · Proceedings · 2014
Summary The present work has as main goal to evaluate alternative code for the GSLIB library to improve the computational performance of the KT3D program for large volumes of data. The approach adopted here was to incorporate OpenMP statement code in KT3D that would allow maximizing the use of available resources in the cards Many Integrated Core (MIC) - Intel Xeon Phi, installed on servers with Intel Xeon CPUs. The results of the study are expected to demonstrate improved times in processing KT3D program running on high-performance processing environment, and show their importance when using other GSLIB applications, for instance the sequential simulations.