“…The average improvement is 1.1%, 9.0%, 15.1%, and 17.2% for two, four, eight, and sixteen threads, respectively. Therefore, taking advantage of Itanium2 caches on FINISTERRAE is critical, because accesses to memory incur in a high penalty in comparison with other systems [11]. Table 3 shows the SpMV performance without optimizations and when the optimizations detailed previously (data and thread allocation, and locality improvement) are applied together.…”