home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.benchmarks:1644 comp.arch:10608
- Newsgroups: comp.benchmarks,comp.arch
- Path: sparky!uunet!ferkel.ucsb.edu!taco!rock!stanford.edu!ames!decwrl!pa.dec.com!engage.pko.dec.com!adodem.enet.dec.com!sharpe
- From: sharpe@adodem.enet.dec.com
- Subject: Re: DEC ALPHA Performance Claims
- Message-ID: <1992Nov11.043149.17257@engage.pko.dec.com>
- Sender: newsdaemon@engage.pko.dec.com (USENET News Daemon)
- Organization: Digital Equipment Corporation
- Date: Wed, 11 Nov 1992 19:51:31 GMT
- Lines: 38
-
-
- In article <BxJ47t.K65.2@cs.cmu.edu>, lindsay+@cs.cmu.edu (Donald Lindsay) writes...
- >
- >zapantis@uvphys.phys.UVic.CA writes:
- >>How does the SPECfp92 rating relates to Mflops? Is it a one-to-one
- >>relationship?
- >>If not, have you rated the Alpha systems( all of them!) in Mflops yet?
- >>care to publish the rating?
- >
- >There are lots of ways to measure MFLOPS, and they all give different
- >answers. Which one are you requesting?
- >
- >At a guess, you're asking for Linpack MFLOPS. This correlates well
- >with *part* of the SPECfp92 test suite. What you would really like is
- >a copy of the "performance brief" which DEC would typically be
- >publishing about now. I expect it will give the Alpha's speed on each
- >individual member of the SPEC suite, plus descriptions.
- >--
- >Don D.C.Lindsay Carnegie Mellon Computer Science
-
- From my copy of EB-N0104-51 "Alpha AXP Server Family Performance Brief -
- OpenVMS", Page 17,
-
- System 100x100 dp 1000x1000 dp
-
- DEC 3000/400S AXP 26.4 70.8
- DEC 3000/500S AXP 30.2 79.9
- DEC 4000/610 AXP 36.3 86.4
- DEC 7000/610 AXP 38.6 102.1
- DEC 10000/610 AXP 42.5 111.6
-
- I don't think I have made any typos.
-
- Dileep will know the reasons for this or that of course :-)
-
- Regards
- Richard Sharpe
- These opinions be mine, Digital has its own!
-