From: pyeatt@Texaco.com (Larry D. Pyeatt) Subject: Re: Coherent vs. Linux - a comparo Date: Tue, 22 Jun 1993 15:12:05 GMT
In article <206rpiINN825@rave.larc.nasa.gov>, jcburt@gatsibm.larc.nasa.gov () writes:
|> You see...there's the rub...I have a 486/33 system, which *new* a year ago
|> cost under $3000. I also use a Sun SparcStation 1+. I've benchmarked both
|> of these "Unix" systems using BYTE's UNIX Benchmark Suite of tests. Both
|> in multi-user mode with 1 user. The amazing thing was, the 486 came out
|> on top in most of the tests. In benchmarking X performance, the 486 again
|> came out on top. My basic conclusion was that the 486 system was *at least*
|> as powerful as an SS1+ computationally. An *additional* plus is that if something
|> goes wrong with the 486 I just pop down to my local computer store and
|> buy the parts...haven't seen too many Sun parts around this area...
|>
|> Now the question...*why* should I buy a *used* Sun, when I can buy a *new*
|> 486 for *less* money *and* be assured that I have access to spare parts...
|> Somehow it just doesn't make sense...
Good question. A 486/33 provides similar integer performance to an older
sparcstation. However, I think the sparcstation will give better floating
performance. I'm not sure about that because the older sparcstations were
really pathetic when it comes to floating point. Does anyone know for
sure? Also, I heard something about a clock doubled sparc chip which can
be inserted into an older sparcstation, almost doubling its processing
power. For floating point, that would definitely beat a 486/66. Integer
performance would be about the same (just a guess).
--
Larry D. Pyeatt This article does not reflect the views
( pronounced "Johnson" ) of my employer or of myself. Any simi-
Internet : pyeatt@texaco.com larity to the views of anyone, real or
Voice : (713) 975-4056 fictional, is purely coincidental.