home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!ames!sun-barr!news2me.EBay.Sun.COM!exodus.Eng.Sun.COM!zarquon!walter
- From: walter@zarquon.Eng.Sun.COM (Walter Bays)
- Newsgroups: comp.benchmarks
- Subject: Re: SPEC : questions, copyright, summary - a database of spec's ?
- Date: 18 Nov 1992 17:52:04 GMT
- Organization: Sun Microsystems
- Lines: 37
- Distribution: world
- Message-ID: <lgl0m4INNp40@exodus.Eng.Sun.COM>
- References: <1992Nov17.191610.6290@news.acns.nwu.edu>
- Reply-To: walter@zarquon.Eng.Sun.COM
- NNTP-Posting-Host: zarquon
-
- >In article <1992Nov16.143948.27382@crc.ac.uk> ifenton@crc.ac.uk (I. Fenton) writes:
- >>i would like to complie a list of all (!) SPECmark ratings, both INT & FP,
-
- SPECint92 and SPECfp92 are the best "single number" metrics for processor speed.
-
- SPECrate_int92 and SPECrate_fp92 are the best "single number" metrics
- for total CPU system throughput. They do not indicate performance of a
- single job, nor of a mixed CPU & I/O workload.
-
- In article 6290@news.acns.nwu.edu, gof@casbah.acns.nwu.edu (Gerald Fountain) writes:
- >This sounds great. I think this would be a great idea. I would vote that you
- >include the component scores in the SPEC ratings as well. They sometimes show
- >some interesting features (HP 700 series has *very* high MATRIX operations it
- >would seem!).
-
- Looking also at individual benchmark SPEC ratios is better than only
- looking at geometric mean, and would help you spot unrepresentative
- anomalies like 030.matrix300 in the old SPEC89 suite. But from my own
- experience, I suspect anyone will find daunting the effort of keeping
- even the summary numbers accurate and up to date.
-
- I would caution anyone maintaining a summary list to be sure to record
- the source and date of each entry, e.g. "Dec-92 Newsletter", "May-92
- Press Release", else it will prove difficult to verify information and
- correct out of date information.
-
- Beginning in January, Release 1 tapes (SPECmark89) will no longer be
- supported, and results from that suite will be marked "obsolete".
-
-
- (Though mostly in accord with my SPEC colleagues, these views nevertheless
- contain my own opinions.)
- ---
- Walter Bays walter.bays@eng.sun.com
- Sun Microsystems, 2550 Garcia Ave., MTV15-404, Mountain View, CA 94043
- (415) 336-3689 SPEC Steering Committee Chairman FAX (415) 968-4873
-
-