home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!spool.mu.edu!think.com!enterpoop.mit.edu!bloom-picayune.mit.edu!athena.mit.edu!jfc
- From: jfc@athena.mit.edu (John F Carr)
- Newsgroups: comp.benchmarks
- Subject: Re: SPECmarks....
- Message-ID: <1992Dec27.174152.16272@athena.mit.edu>
- Date: 27 Dec 92 17:41:52 GMT
- References: <1992Dec24.220732.11056@zooid.guild.org> <PHR.92Dec25152236@napa.telebit.com>
- Sender: news@athena.mit.edu (News system)
- Organization: Massachusetts Institute of Technology
- Lines: 19
- Nntp-Posting-Host: achates.mit.edu
-
- In article <PHR.92Dec25152236@napa.telebit.com>
- phr@telebit.com (Paul Rubin) writes:
-
- >But nowadays
- >usually 1 specmark = about 1.5 mips according to most people's
- >measurement of mips. This is peculiar because a VAX 11/780,
- >supposedly a 1 mip machine, runs (by definition) 1.00000 specmarks.
-
- A VAX 11/780 is closer to 1.2 specmarks. A specmark is defined as a 780
- with a particular set of compilers. Newer compilers are better.
-
- If you look at individual SPEC benchmarks, you'll find 50% differences in
- the ratios for some machines. I think the DEC Alpha ranges from 60 to 100
- on the integer programs. Why should the dhrystone test correlate any more
- with the SPEC average than the SPEC tests themselves do?
-
-
- --
- John Carr (jfc@athena.mit.edu)
-