hedley%wnre.aecl.cdn%ubc.CSNET@RELAY.CS.NET (Bruce Hedley) (03/07/87)
Our group at Whiteshell are using the Harwell Sparse Matrix Package as the core of our CANDU loss-of-coolant-accident analysis code, which is currently running on a VAXCluster (1-8650,1-785,2-750s). We have recently had a FLoating Point Systems FPS-364 linked to the cluster, and it runs essentially as a fast batch queue. I have been looking around to see if a vectorised version of the Harwell code exists, with no luck. FPS assures me that none exists (perhaps in hope of getting a contract to do it!), but I have heard a story that such a version exists out on the US West Coast. If this is true, or if anyone has heard words to that effect, I would greatly appreciate hearing from them, either via E-mail or the Alexander Graham Bell. Thanks in advance to any and all who reply..... Bruce Hedley Research Analyst Atomic Energy of Canada Limited Whiteshell Nuclear Research Establishment Pinawa, Manitoba, CANADA R0E 1L0 (204)-753-2311-2700