U.C. Berkeley Millennium: Workstation Cluster For Parallel Distributed Computing In Mechanical Engineering

Progess Reports

1998 Q1 The single four-processor Intel machine has been set-up in the Davis Etcheverry Computing Facility (DECF) where it is being jointly managed by systems administrators from DECF, IEOR and Nuclear Engineering. A steering committee with members from ME, IEOR and NE has been formed to provide administrative guidance for the cluster to be shared among the three departments. Specifically, this group will coordinate installation of new machines and the networking in the cluster as the hardware becomes available.

To this point, use of the single machine received from Intel has been very light because there has been little software available for it. There was some initial difficulty in setting up even basic C and Fortran compilers fo the Solaris OS. Now they are in place and some investigators are trying-out some of their more demanding code onthenew machine. At this point we do not have software that will permit us to run multithread programs that will use all four processors. When software with this capability become available, and we get another machine for the ME/NE/IEOR NOW, we expect the useage level to increase.

1998 Q2
February 1999