mark@mips.COM (Mark G. Johnson) (02/12/91)
Here are all of the accumulated measurements on the pi program.
DEC VAX-9000 (VMS 5.4) 0.39 seconds mk@uta.fi
IBM 3090-600S/VF (AIX370) 0.8 seconds KXM@PSUVM.PSU.EDU
IBM RS/6000 model 540 1.15 seconds rs0xeg@rohmhaas.com
IBM RS/6000 model 530 1.3 seconds chriss@denali.ece.cmu.edu
IBM RS/6000 model 730 1.35 seconds shair@ux1.cso.uiuc.edu
IBM RS/6000 model 320 1.68 seconds twood@UCSD.EDU
IBM RS/6000 model 520 1.74 seconds shair@ux1.cso.uiuc.edu
SGI Personal Iris 4D/35 1.8 seconds wiltse@sgi.com
Cray Y/MP 2.67 seconds mccalpin@perelandra.cms.udel.edu
MIPS Magnum workstation 2.8 seconds mark@mips.com
DEC DECstation 5000/200CX 3.0 seconds terry@venus.sunquest.com
Evans & Sutherland ESV/3 3.6 seconds rthomson@dsd.es.com
SGI Personal Iris 4D/25 3.7 seconds wiltse@sgi.com
Cray 2S 4.15 seconds KXM@PSUVM.PSU.EDU
HP 9000/400t 4.6 seconds dmnewton@UCSD.EDU
NeXTstation (68040) 4.8 seconds andy@juno.caltech.edu
Compaq 386/33 5.1 seconds suitti@ima.isc.com
Sun SPARCstation 2 5.1 seconds mawa@sun1.ruf.uni-freiburg.de
Compaq 486/25 5.1 seconds suitti@ima.isc.com
DEC DECstation 2100 5.5 seconds codex!peterd@uunet.UU.NET
Solbourne S4000 6.2 seconds jmoore@Solbourne.COM
Compaq 386/25 6.8 seconds suitti@ima.isc.com
Sun SPARCserver 4/330 8.5 seconds alan@frey.newcastle.edu.au
Sun SPARCstation 1+ 9.5 seconds mark@mips.com
Apple MacIntosh MacII ci 10.0 seconds torrie@cs.stanford.edu
Sun 3/60 11.7 seconds robp@anubis.network.com
Sun SPARCstation 1 11.8 seconds mark@mips.com
MEiKO Surf/Transputer T800 13.49 seconds rdd@ukc.ac.uk
If I might be permitted to editorialize, the *relative* performance
of the machines is quite unexpected, meaning that this benchmark
is probably unrepresentative of typical machine loads.
(I mean, do you really think a VAX-9000 is 6.8X faster than a Cray Y/MP?)
NeXT 040 is 1.1X faster than a Cray 2S?
Mac II is faster than a SPARCstation 1?
So, in conclusion, forget this benchmark :-). Although fun, it's misleading.
--
-- Mark Johnson
MIPS Computer Systems, 930 E. Arques M/S 2-02, Sunnyvale, CA 94086
(408) 524-8308 mark@mips.com {or ...!decwrl!mips!mark}aburto@marlin.NOSC.MIL (Alfred A. Aburto) (02/13/91)
In article <45783@mips.mips.COM> mark@mips.COM (Mark G. Johnson) writes: > >If I might be permitted to editorialize, the *relative* performance >of the machines is quite unexpected, meaning that this benchmark >is probably unrepresentative of typical machine loads. > > (I mean, do you really think a VAX-9000 is 6.8X faster than a Cray Y/MP?) > NeXT 040 is 1.1X faster than a Cray 2S? > Mac II is faster than a SPARCstation 1? > > >So, in conclusion, forget this benchmark :-). Although fun, it's misleading. >-- > -- Mark Johnson > MIPS Computer Systems, 930 E. Arques M/S 2-02, Sunnyvale, CA 94086 > (408) 524-8308 mark@mips.com {or ...!decwrl!mips!mark} ---------- Mark, Don't give up yet! Here are some more results to consider: VAX 8600 4.3 BSD UNIX (gcc -O) 4.2 sec VAX 8600 4.3 BSD UNIX ( cc -O) 4.5 sec Amiga 2000 (68030/33MHz, ICache ON, DCache ON, Burst OFF) 9.6 sec Amiga 2000 (68030/25MHz, ICache ON, DCache ON, Burst OFF) 12.3 sec Amiga 2000 (68030/25MHz, ICache OFF, DCache OFF, Burst OFF) 14.3 sec Grid LapTop 386/12.5MHz (Turbo C V2.0) 173.6 sec Zenith Z-248 286/ 8.0MHz (Turbo C V2.0) 282.4 sec The results may not represent (ahem) 'typical' usage, but they are real results after all (as long as some didn't extraordinarily tweak the code). Al Aburto aburto@marlin.nosc.mil
co@fy.chalmers.se (Christer Olsson) (02/14/91)
I've missed the pi benchmark because some problems at our news-machine. Can anybody mail me an example of the program? (I want to test it on my _slow_ 68008-based Xenix-machine.. :-) ) co@adagio.fy.chalmers.se or co@chrom1.mednet.gu.se