NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/comp/os/linux/16197

Internet Message Format | 1992-11-09 | 3.5 KB

Xref: sparky comp.os.linux:16197 comp.sys.ibm.pc.hardware:29355
Path: sparky!uunet!ferkel.ucsb.edu!taco!rock!stanford.edu!ames!elroy.jpl.nasa.gov!news.claremont.edu!ucivax!news.service.uci.edu!beckman.com!daparish
From: [e-mail]
Newsgroups: comp.os.linux,comp.sys.ibm.pc.hardware
Subject: Re: 486SXs as Unix Iron?
Message-ID: <[e-mail]>
Date: 9 Nov 92 16:47:54 PDT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Organization: Beckman Instruments, Inc.
Lines: 58

In article <[e-mail]>, [e-mail] (fred j mccall 575-3539) writes:
> In <[e-mail]> [e-mail] (peter da silva) writes:
>
>>In article <[e-mail]> [e-mail] (Tim Pierce) writes:
>>> I'm looking for a machine on which to run Linux in the $1500-$2000
>>> range.  I was looking at a lot of 486SX boxes, until I ran across Eric
>>> Raymond's PC Unix hardware buyer's guide, which includes the following
>>> paragraph:
>
>>> > Don't bother with SX machines.  Under UNIX the 16-bit bus-to-CPU path
>>> > can choke your throughput.

Obviously this was referring to 386sx machines, since 486sx have same data
path as 486dx.

>>> > The 486SX is even worse, a stupid marketing crock
>>> > with no technical justification whatsoever.
>
>>This is Eric's politics coming to the fore. A 486SX is a better choice than
>>any 386DX I know of. The 487SX is a scam, but most competantly designed
>>motherboards support the 486SX or DX with a jumper change.
>
>>Eric's a nice guy, he's just got less tolerance for marketing bullshit than
>>most and it shows. Objectively, the 486SX is an OK chip, and the marketing
>>bullshit can be avoided by careful motherboard choices.
>
> Yeah, but on this one I agree with him.  The only way to keep
> companies from doing silly marketing cruft like this is if people let
> them eat all those chips.  There was NO reason to come out with the
> 486SX other than to attempt predatory pricing on folks like AMD and
> Cyrix while keeping the price of the 486DX inflated.  There is no way
> I'm going to believe that the price difference between the 486SX and
> the 486DX can be explained by the cost of testing the math unit in the
> latter.

Marketing hype, be danged! The 486sx33 looks like a bargain to me.
So why not take advantage of the artificially low price of the 486sx?
(or is the 486dx33 thats artificially high?)
It performs like a 486dx, and such a small percentage of machines (<10%?)
ever use the fpu anyway (only CAD users and a few esoteric scientific apps
perhaps?). Other than CAD there is no evidence that any of the major commercial
apps, not Unix (except maybe X), not even Windows NT, will take take advantage
of an fpu.
Hey, I'm a 486dx33 owner, but as far as I know the fpu's never been used. If I
had to by the machine over I'd buy a 486sx33 and spend the other $200+
elsewhere.
486sx/33's should cost about $100 over a 386dx33 with twice the performance.
Chances are if you ever need the fpu, prices will be way down (on those old,
slow 33mhz chips). Hell, you'll
probably be able to by a 486dx33 for the delta cost if you have too. Just get
a MB that takes either 486sx or dx and you'll be ok.

--
Ashley parish    [e-mail]
   .___                /In Oklahoma city, enthusiastic meter maids
     o/                   /slapped 12 tickets on a car illegally parked
     |                  /for 3 days. They failed to notice the driver,
     >\                 /after a heart attack, dead behind the wheel.