Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!ncar!tank!uxc!uxc.cso.uiuc.edu!uxg.cso.uiuc.edu!phil
From: phil@uxg.cso.uiuc.edu
Newsgroups: comp.arch
Subject: Re: Quadruple-Precision Floating Point
Message-ID: <400001@uxg.cso.uiuc.edu>
Date: 20 Dec 88 20:13:00 GMT
References: <8561@alice.UUCP>
Lines: 11
Nf-ID: #R:alice.UUCP:8561:uxg.cso.uiuc.edu:400001:000:539
Nf-From: uxg.cso.uiuc.edu!phil    Dec 20 14:13:00 1988


For TRIVIAL extended precision work, IBM VM/CMS Rexx language has N-digit
precision fixed and floating point numbers, but NO function library.  It
is SLOOOOOW, but it is handy for those few small calculations needing an
infinite precision.  I have implemented a SQRT function with variable
precision under Rexx.

I wish some machine had truly VARIABLE precision firmware floating AND FIXED
point (i.e with fractions) arithmetic.  On a RISC machine, an architecture
that allowed easy extension of precision would be just as good.

--phil--