Xref: utzoo comp.sys.sgi:6653 comp.arch:18999 Path: utzoo!attcan!uunet!samsung!uakari.primate.wisc.edu!uflorida!travis!tom From: tom@ssd.csd.harris.com (Tom Horsley) Newsgroups: comp.sys.sgi,comp.arch Subject: Re: SPECthruput Message-ID: Date: 5 Nov 90 11:57:58 GMT References: <42673@mips.mips.COM> <1990Nov4.022808.15241@csusac.csus.edu> Sender: news@travis.csd.harris.com Followup-To: comp.sys.sgi Organization: Harris Computer Systems Division Lines: 25 In-reply-to: croft@csusac.csus.edu's message of 4 Nov 90 02:28:08 GMT >>>>> Regarding Re: SPECthruput; croft@csusac.csus.edu (Steve Croft) adds: croft> Is there really a SPEC benchmark called 'SPECthuput'? I have seen this croft> mark referred to in different manufacturers' literature. Yep. They (SPEC that is) defined this is one of their newsletters a while back. Basically it consists of running 2 copies of the regular SPEC suite at the same time on each CPU in your system (so if you have a 4 CPU multiprocessor, you would be running a total of 8 copes of SPEC). You measure the wall time from the time the first copy starts to the time the last copy finishes and use that figure to compute SPECthroughput in much the same way you use wall time for a single copy to compute SPECmarks. The numbers are supposed to be reported in the format 4@999.9 which means you came up with a SPECthroughput number of 999.9 on a 4 CPU system. (Before anyone asks, this is not a real number from some system I have seen). -- ====================================================================== domain: tahorsley@csd.harris.com USMail: Tom Horsley uucp: ...!uunet!hcx1!tahorsley 511 Kingbird Circle Delray Beach, FL 33444 +==== Censorship is the only form of Obscenity ======================+ | (Wait, I forgot government tobacco subsidies...) | +====================================================================+