Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: comp.arch
Subject: Re: Floating point accuracy
Message-ID: <7313@utzoo.UUCP>
Date: Tue, 11-Nov-86 14:40:29 EST
Article-I.D.: utzoo.7313
Posted: Tue Nov 11 14:40:29 1986
Date-Received: Tue, 11-Nov-86 14:40:29 EST
References: <340@euroies.UUCP> <1989@videovax.UUCP>, <267@bath63.UUCP>
Organization: U of Toronto Zoology
Lines: 13

> The mechanism was to add a random number in the range 0 to 1 in the last
> digit and then truncate. This gives better RMS error than most other rounding
> schemes. It also has the advantage that numerically unstable programs are
> easy to detect (they give wildly differing results on each run).
> Sadly I believe this has never been implemented...

The IBM 7030, aka Stretch, had a random-rounding mode, although possibly
using a somewhat simpler method.  The idea was to run your program twice,
once with normal rounding and once with random rounding, and see how the
results differed.
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,decvax,pyramid}!utzoo!henry