Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1997
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1997

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Pentium bug

  • To: mathgroup at smc.vnet.net
  • Subject: [mg7226] Re: [mg7209] Pentium bug
  • From: jpk at max.mpae.gwdg.de
  • Date: Sat, 17 May 1997 19:13:13 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

> From cmorpurgo at mail.utexas.edu Fri May 16 10:03:00 1997
> Date: Fri, 16 May 1997 02:30:48 -0400 (EDT)
> From: cmorpurgo at mail.utexas.edu (Carlo Morpurgo)
> To: mathgroup at smc.vnet.net
> Subject: [mg7209] Pentium bug
> 
>  Dear all, I am about to buy  a Pentium pro 200 - 
>  Recently Intel has admitted bugs in such chips. I wonder
>  if anyone knows about the possible impacts on Mathematica
>  computations.
> 
>  Thanks, Carlo
> 
My favored computer magazine says that the bug is related to the 
floating point exceptions of the FPU. Microsofts C Compiler
usualy ignore the exceptions completly. Wiht Borland C You can
mask the exceptions easily. Probably Mma 3.0 for Windows is compiled
with Microsoft C x.x so this does not matter.

More over since Mma does Integer arithmetic and a lot of floting point
calculation using it's own multiprecission 
arithmetic I expect no problems.

Finaly AMD K6 has the error not and it is at least cheeper to get an AMD
chip.

Hope that helps 
Jens


  • Prev by Date: Re: How to clear all definitions?
  • Next by Date: Re: Until -Reply -Reply
  • Previous by thread: Pentium bug
  • Next by thread: Palettes!! with MMA 3.0