Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site alice.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!alice!ark
From: ark@alice.UucP (Andrew Koenig)
Newsgroups: net.arch
Subject: Re: Oh no! More integer division
Message-ID: <5059@alice.uUCp>
Date: Fri, 28-Feb-86 17:17:45 EST
Article-I.D.: alice.5059
Posted: Fri Feb 28 17:17:45 1986
Date-Received: Sat, 1-Mar-86 17:46:07 EST
References: <5100014@ccvaxa>
Organization: Bell Labs, Murray Hill
Lines: 11

> Now, is this an architectural question: should or could randomized rounding be
> provided by a floating point unit? Assuming you can cheaply obtain a random
> number in hardware (amplify transistor noise?) doing randomized rounding in
> the floating point unit would be a lot faster than doing so in software,
> where you'd have to extract and play around with the guard bits.

If randomized rounding is provided by a floating-point unit,
it had better be easy to turn off because using it is likely
to make debugging impossible.  Consider: if you run the same
program twice and get slightly different results, how will you
ever know if it is a bug in the program?