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

80 bit floating point math

• To: mathgroup at smc.vnet.net
• Subject: [mg4521] 80 bit floating point math
• From: emoore at acm.org (Eric Moore)
• Date: Wed, 7 Aug 1996 04:17:32 -0400
• Organization: Compumedia Internet Services
• Sender: owner-wri-mathgroup at wolfram.com

I am doing a test at Digital Corporation.  I know that the win16 api 
x86 emulator that comes with windows NT 4.0 has problems with "temp real" or 
80 bit floating point numbers.  I now want to test digital's fx!32 win32 api 
x86 emulator's ability to handle these type of numbers.

I figured the best way to do that would be to use Mathematica,

I have discovered the N function but I don't see how I can guarantee myself 
that this digit precision mechanism would ever create a 80 bit precison 
floating point number.

I printed the MAX(something) variable and after a quick convesion I found out 
that it was 2^1024 or another words a long real (63 bit precision).

please reply to moore at ntmail.zso.digital.com as I don't normally frequent this 
nesw group.

==== [MESSAGE SEPARATOR] ====