High-End Computing on SHV Workstations Connected with High Performance Network

In this paper, we present performance of a set of typically compute intensive benchmarks on a 16 node SCI based PC cluster. The performance is compared with other currently used high-end computers. Dedicated clusters with high-performance interconnect are shown to be competitive in absolute performance to more traditional high-end computers, offering good scalable performance at a low price.

[1]  Message P Forum,et al.  MPI: A Message-Passing Interface Standard , 1994 .

[2]  Gordon E. Moore,et al.  Progress in digital integrated electronics , 1975 .

[3]  Lars Paul Huse,et al.  ScaMPI-Design and Implementation , 1999, Scalable Coherent Interface.

[4]  Remzi H. Arpaci-Dusseau,et al.  Architectural Requirements and Scalability of the NAS Parallel Benchmarks , 1999, ACM/IEEE SC 1999 Conference (SC'99).

[5]  G.E. Moore,et al.  Cramming More Components Onto Integrated Circuits , 1998, Proceedings of the IEEE.

[6]  Hakon O. Bugge,et al.  Affordable Scalability Using Multi-Cubes , 1999, Scalable Coherent Interface.