home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.arch
- Path: sparky!uunet!decwrl!pa.dec.com!rdg.dec.com!news.crl.dec.com!jg
- From: jg@crl.dec.com (Jim Gettys)
- Subject: Re: DEC Alpha AXP System Performance
- Message-ID: <1992Nov12.101520.2828@crl.dec.com>
- Keywords: Alpha, AXP, SPEC, DEC
- Sender: news@crl.dec.com (USENET News System)
- Organization: DEC Cambridge Research Lab
- References: <1992Nov10.153629.27510@ryn.mro4.dec.com> <martin.721554717@bert>
- Date: Thu, 12 Nov 1992 10:15:20 GMT
- Lines: 39
-
- In article <martin.721554717@bert>, martin@math.rwth-aachen.de ( Martin Schoenert) writes:
- > DEC 3000 Model 500 AXP, @150 MHz
- > SPECint92 74.30
- > SPECfp92 126.0
- >
- > DEC 3000 Model 400 AXP, @133 MHz
- > SPECint92 65.32
- > SPECfp92 112.2
- >
- > In front of me I have Digital Equiment Germany's announcement of the new
- > systems for ``Forschung und Lehre'' (research and education). For one
- > thing it doesn't mention the first three models at all. Also it gives
- > the following numbers
- >
- > DEC 3000 AXP Model 500
- > SPECint92 66
- > SPECfp92 118
- >
- > DEC 3000 AXP Model 400
- > SPECint92 59
- > SPECfp92 103,5
- >
- > Any idea why the numbers are smaller? (Is it because we are ``Forschung
- > und Lehre'', or is it because we are in Germany, or what? ;-)
-
- The reason is simple: folks were working on the compilers and remeasuring SPEC results
- up until the last possible moment before annoucement. Some earlier results
- seem to have slipped through the cracks. Of course, we continue to work
- on the compilers, and yet better results can be expected in the future :-).
- Significant improvements were made in even the last week (and weekend) before
- annoucement.
-
- In any case, Dileep's numbers are the first set
- of "official" correct ones and were submitted to SPEC.
- - Jim
-
- --
- Digital Equipment Corporation
- Cambridge Research Laboratory
-