NCSA

SGI Power Challenge (distributed shared memory)

6 Power Challenge machines (5 with 16 processors, 1 with 8

1 server with 12)

64 bit word

4MB cache per processor

Compilers: F77, F90, C, C++

Message Passing: MPI, SCI’s MPI, PVM, HPF

HP-Convex Examplar SPP-1200

8 HYPERnodes with 8 processors

shared memory within HYPERnodes

Software: f77, f90,C, C++

Message Passing: MPI, Convex PVM

Silicon Graphics CRAY Origin2000- in friendly user mode

128 processors

distributed shared memory

Compilers:f77, f90,C,C++,HPF

Message Passing: SGI native MPI, PVM

HP-Convex Examplar S-Class - not yet available for use

16 processors

Previous slide

Next slide

Back to first slide

View graphic version