Porovnaní vybraných vývojových platforem z hlediska výkonu v grafických aplikacích

Abstract

The aim of this work is to select an appropriate methodology for measuring the time for a critical section of code, used in graphics applications. We also need to select the appropriate algorithm and demonstrate it to the individual optimization techniques such as SSE instructions or OpenMP directives and TBB libraries to parallelize code. Part of this work is a set of recommendations including advice and demonstrations for code optimization using performance monitor counters and assembly language.