1k1mgm@kuhub.cc.ukans.edu (Christopher Gunn) (11/17/90)
Word-of-mouth benchmark results I've received suggests that 4D/35 is about 10% integer/20% f.p. FASTER than 4D/310. If this is due to 4D/35 memory architecture, I'll take back some (not all) of the dark thoughts I had upon learning that 4D/35 SIMMs used non-standard architecture. On the other hand, is it true? I've got SPEC results for 4D/320 (SPEC jobs on one CPU) but no Dongarra/Linpack stuff for either 3x0 or 35, and no SPEC stuff for 35. Does anybody know where I can get this data? [I realize benchmarks are generally crapola, but this information would be helpful for some grant development. We got good feedback from NIH for beating this issue right into the ground on recent proposal (still pending for real $$$).] Christopher Gunn Molecular Graphics and Modeling Lab SPAN--KUPHSX::GUNN Department of Medicinal Chemistry, Malott Hall 913-864-4428 or -4495 University of Kansas, Lawrence, KS 66045