pb@idca.tds.PHILIPS.nl (Peter Brouwer) (09/14/90)
We are searching for benchmarks that can be used to characterize the behaviour of multi processor UNIX implementations. Are there such benchmarks ? -- # Peter Brouwer, | Philips Information Systems, # # NET : pb@idca.tds.philips.nl | Department P9000-i Building V2, # # UUCP : ....!mcsun!philapd!pb | P.O.Box 245,7300AE Apeldoorn,The Netherlands# # PHONE:ext [+31] [0]55 432992, | FAX :ext [+31] [0]55 433488 #
walter@vogons.UUCP (Walter Bays) (09/28/90)
In article <637@ssp2.idca.tds.philips.nl> pb@idca.tds.PHILIPS.nl (Peter Brouwer) writes: >We are searching for benchmarks that can be used to characterize the behaviour >of multi processor UNIX implementations. There are a number of benchmarks, commercial and non-commercial, that attempt to characterize system performance versus number of processes or number of simulated users for various workloads. There are famous cases of poor benchmark design and misinterpretation of results in this area, so be sure you understand your workload and look carefully at the benchmark(s) you use. The "Thruput Method A" of the SPEC Release 1.1 benchmarks can be used to measure the additional throughput provided by multiple processors for an homogenous workload. If your workload is CPU-bound, this would make a good starting point, because the release 1.1 benchmarks are CPU benchmarks. SPEC is currently considering several I/O and system-level benchmarks for Release 2.0. For more information, you can call SPEC at (415) 792-3334. Or if your workload is transaction processing you could consider the TPC benchmarks. --- My_own_opinions_disclaimer + Cute_saying Walter Bays Phone (415) 852-2384 FAX (415) 856-9224 EMAIL uunet!ingr.com!bays or uunet!{apple.com,pyramid.com}!garth!walter USPS: Intergraph APD, 2400 Geng Road, Palo Alto, California 94303