Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!ames!amdahl!dgcad!dg-rtp!webo!cheshirecat!lewine From: lewine@cheshirecat.webo.dg.com (Donald Lewine) Newsgroups: comp.sys.m88k Subject: Re: Data General Aviion 7000, 8000 Message-ID: <1991Mar27.135031.29533@webo.dg.com> Date: 27 Mar 91 13:50:31 GMT References: <1991Mar14.183029.11714@neon.Stanford.EDU> Sender: usenet@webo.dg.com (Usenet Administration) Reply-To: uunet!dg!lewine Distribution: comp Organization: Data General Corporation Lines: 31 In article , deraadt@fsa.cpsc.ucalgary.ca (deraadt) writes: |> In article <1991Mar22.143552.16204@webo.dg.com> lewine@cheshirecat.webo.dg.com |> (Donald Lewine) writes: |> > A more meaningful metric is the SPEC benchmark. The AViiON gets |> > an overall SPECthrughput of 54. It does even better if you leave |> > out the double precision nuclear physics test. |> |> I've been eagerly awaiting a SPEC rating rather than just a SPECthruput. |> As I understand it, the SPECthruput is not just ncpu*SPEC, but some other |> contortion. |> |> Is DG afraid to give plain SPEC ratings, hiding behind SPECthruput? |> The SPECthruput measures a multi-processor. The SPECmark measures a uni-processor. The thruput test runs 2xCPUS copies of the spec tests and measures the total time to run all of the copies. The SPECthruput number is what a real user would see and includes the scheduler overhead and any other interference effects that would affect performance. It is not a "contortion". The SPECmark ratings are not meaningful because 3 of the CPUs would be idle for the test. The SPECthrughput was invented to measure multiprocessor systems. -------------------------------------------------------------------- Donald A. Lewine (508) 870-9008 Voice Data General Corporation (508) 366-0750 FAX 4400 Computer Drive. MS D112A Westboro, MA 01580 U.S.A. uucp: uunet!dg!lewine Internet: lewine@cheshirecat.webo.dg.com