XcalableMP for Productivity and Performance in HPC Challenge Award Competition Class 2

In this paper, we present our XcalableMP implementation of the HPCC HPL, RandomAccess, FFT, and the Himeno benchmark [1] which is a typical stencil application. The highlights of this submission are as follows: • We implemented three HPCC benchmarks; HPL, RandomAccess, and FFT. In addition, we implemented the Himeno benchmark. • The SLOC (Source lines of code) of the benchmarks is shown in Table 1. • The performance summary of the benchmarks is shown in Table 2. • To measure performances, we used the K computer at RIKEN AICS and HA-PACS at University of Tsukuba in Japan (Table 3).