Dávkové zpracování biomedicínských výpočtů na superpočítači Rivyera

Abstract

This thesis describes batch processing of data at hardware architecture of gate arrays represented by Rivyera supercomputer, its computational power is mediated by several dozens of field programmable gate arrays. The aim of thesis is to design and implement solution of remote control and batch processing of data at Rivyera supercomputer and also to provide a method to development and resolution of calculation by using parallel computing of Rivyera supercomputer. Computational methods are solved by combining of computational core of field programmable gate arrays, described using VHDL language, and handler host application, written in Java programming language. An example of computational methods using Rivyera supercomputer is parallel processing of CT images into digitally reconstructed radiographs from a certain directional projection.