Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!bionet!agate!ucbvax!decwrl!decvax!crltrx!jg
From: jg@crltrx.crl.dec.com (Jim Gettys)
Newsgroups: comp.sys.dec
Subject: Re: DEC station 3100
Keywords: DEC performance PMAX DECstation 3100
Message-ID: <105@crltrx.crl.dec.com>
Date: 19 Mar 89 20:38:17 GMT
References: <272@cbnewsl.ATT.COM> <7561@batcomputer.tn.cornell.edu> <3085@stpstn.UUCP> <7579@batcomputer.tn.cornell.edu>
Reply-To: jg@crltrx.crl.dec.com.UUCP (Jim Gettys)
Distribution: usa
Organization: DEC Cambridge Research Lab
Lines: 21

In article <7579@batcomputer.tn.cornell.edu> thompson@tcgould.tn.cornell.edu (Steve Thompson) writes:
>In article <3085@stpstn.UUCP> aad@stpstn.UUCP (Anthony A. Datri) writes:
>>In article <7561@batcomputer.tn.cornell.edu> thompson@tcgould.tn.cornell.edu (Steve Thompson) writes:
>>>compared the performance of my main application (a floating-point
>>>intensive code that runs for several hours) between the DS3100 and
>>>a Sun 4/110. The DS3100 was 3.34 times faster.
>>
>>Does that 4/110 have the optional floating point?  I've found that ours
>>is terrible at floating point because it doesn't.
>
>Yup, the 4/110 has an FPU.

Yes, the PMAX (DECstation 3100) is that fast. :-)  3.34 is well within the
typical performance variance between the Sun 4/110 and the PMAX.  See the
latest MIPS Performance Brief (version 3.6) that MIPS puts out; it has
lots of useful data of many benchmarks; I provided John Mashey with the
PMAX numbers when the machine was announced (John was kind enough to provide
me with a tape of the benchmark suite). 
				Jim Gettys
				Digital Equipment Corporation
				Cambridge Reserch Laboratory