🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.ibm.pc.games.rpg
Article #584

Re: Problem with Pentium processors

#584
From: ak... Allen Kim
Date: Mon, 05 Dec 1994 19:12
41 lines
3246 bytes
cra...@herbie.unl.edu (Chris Radek) writes:

>I'd go along with the fact that it's *not* all that unlikely for an average
>individual to encounter the problem on their Pentium.  I downloaded the
>test program from CompuServe and ran it on my Gateway P5-60 at work.  The
>answer was off by 256 for the numbers given in the previous article:
> 
>    4195835 - ((4195835/3145727)*3145727 = ?
> 
>It should be 0.  My Pentium told me it was 256.  I don't know if these 
>numbers are the only ones which produce an incorrect result, although I 
>would doubt they are unique.  The defense that it takes "billions and 
>billions" of operations involving things like weather simulations, planetary 
>spacecraft trajectories, and other types of rocket-science is laughable.  
>What if these numbers (or others which produce incorrect results) should 
>happen to come up in one of the thousands of calculations my company does 
>for payroll, taxes, etc?  If we under-pay our income taxes by several 
>million dollars, is Intel going to explain this to the IRS, and then pay the 
>(undoubtedly large) penalty which results?  Or, on a more intimate and 
>frightening note, what if a hospital or drug company should happen to hit 
>these numbers in a calculation involving drug dosages and it goes unnoticed?  
>How will Intel explain away all the corpses which might result?  Saying that 
>it is, essentially, "Not a big deal" borders on criminal negligence in my 
>book.

They do have a 1-800 number for you to call if you need a replacement.  Intel
used to ask many questions dealing with your appanent "need" for double-
precision floating-point calculations, but now they aren't asking as much.
If you are going to use your Pentium for any mathematical work now or in the
near future, you shouldn't have any trouble getting them to approve you for a
replacement.

By the way, about your hospital and drug company example: Some of the first
people Intel approved for a replacement were INDEED those working in medical
labs.  Others made the crack about how you should fib that you will be using
your Pentium for medical work if you want to be guaranteed a replacement...

Check out the comp.sys.intel newsgroup.  They've got posters there with
torches, pitchforks, nooses, and picket signs saying "Down with Intel!"

				- Allen "Ninja" Kim
Technical Details
Message-ID: <3bvon8$rur@tuba.cit.cornell.edu>
Path: mbox-mo1.pugleaf.net!bga.com!news.sprintlink.net!howland.reston.ans.net!spool.mu.edu!caen!news.tc.cornell.edu!travelers.mail.cornell.edu!tuba.cit.cornell.edu!crux4!ak10
References: <3b3mqa$jb3@homesick.cs.unlv.edu> <JABRAM.94Nov29154420@pdx025.ichips.intel.com> <pauldb.30.000D7C6B@dsusmail.dsus.datastorm.com> <3bu3mr$t1c@crcnis3.unl.edu>

Show Full Headers
Complete Article Headers
Path: mbox-mo1.pugleaf.net!bga.com!news.sprintlink.net!howland.reston.ans.net!spool.mu.edu!caen!news.tc.cornell.edu!travelers.mail.cornell.edu!tuba.cit.cornell.edu!crux4!ak10
From: ak...@crux4.cit.cornell.edu (Allen Kim)
Newsgroups: comp.sys.ibm.pc.games.action,comp.sys.ibm.pc.games.rpg,comp.sys.ibm.pc.games.adventure,comp.sys.ibm.pc.strategic,comp.sys.ibm.pc.games.marketplace
Subject: Re: Problem with Pentium processors
Date: 5 Dec 1994 19:12:40 GMT
Organization: Cornell University
Lines: 41
Message-ID: <3bvon8$rur@tuba.cit.cornell.edu>
References: <3b3mqa$jb3@homesick.cs.unlv.edu> <JABRAM.94Nov29154420@pdx025.ichips.intel.com> <pauldb.30.000D7C6B@dsusmail.dsus.datastorm.com> <3bu3mr$t1c@crcnis3.unl.edu>
NNTP-Posting-Host: 128.253.232.66
X-Newsreader: NN version 6.5.0 #6 (NOV)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit