Architecture of the VPP500 parallel supercomputer
Proceedings of the 1994 ACM/IEEE conference on Supercomputing
LAPACK Working Note 43: A Look at Scalable Dense Linear Algebra Libraries
LAPACK Working Note 43: A Look at Scalable Dense Linear Algebra Libraries
Scalar processor of the VPP500 parallel supercomputer
ICS '95 Proceedings of the 9th international conference on Supercomputing
Development of a mathematical subroutine library for Fujitsu vector parallel processors
ICS '98 Proceedings of the 12th international conference on Supercomputing
Architecture of the VPP500 parallel supercomputer
Proceedings of the 1994 ACM/IEEE conference on Supercomputing
Hi-index | 0.00 |
This paper describes the implementation of two high performance linear equation solvers developed for the Fujitsu VPP500, a distributed memory parallel supercomputer system. The solvers take advantage of the key architectural features of VPP500--(1) scalability for an arbitrary number of processors up to 222 processors, (2) flexible data transfer among processors provided by a crossbar interconnection network, (3) vector processing capability on each processor, and (4) overlapped computation and data transfer. The general linear equation solver based on the blocked LU decomposition method achieves 120.0 GFLOPS performance with 100 processors in the LINPACK Highly Parallel Computing benchmark.