Looking toward Exascale Computing

DOI: 10.1109/PDCAT.2008.88

关键词: Computer architecture 、 Multi-core processor 、 Architecture 、 Supercomputer 、 Computer science 、 Focus (computing) 、 Exascale computing 、 Fault management 、 Operating system 、 System software 、 Electric power

摘要: Recently, Argonne National Laboratory installed a half petaflop Blue Gene/P system. It is the world's fastest open science supercomputer. With 163,840 cores, machine beginning to provide insight on how we might build future platforms as scale toward exascale computing. There are many challenges, including dramatic shift multicore, cost of electric power, and need for robust fault management. In this talk I will focus architecture system software challenges face continue attack ever-larger computational problems.