Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!decwrl!sgi!shinobu!rodman From: rodman@sgi.com (Paul K. Rodman) Newsgroups: comp.arch Subject: Re: Implementing Interval Arithmetic with IEEE rounding modes Message-ID: <1991Jun19.165150.2121@shinobu.sgi.com> Date: 19 Jun 91 16:51:50 GMT References: <9106190449.AA02871@ucbvax.Berkeley.EDU> Sender: news@shinobu.sgi.com (Net News) Organization: Silicon Graphics, Inc., Mountain View, CA Lines: 35 In article <9106190449.AA02871@ucbvax.Berkeley.EDU> jbs@WATSON.IBM.COM writes: > Robert Herdon asked: >While I've seen lots of verbage on the pros and cons of interval.. I've heard yapping about interval arithmetic for a decade and have yet to see anyone that actually _used_ it. Has anyone implemented a compiler with set of "interval" data types? Has anyone used interval arithmetic by doing it themselves with macros/functions/subroutines. I'd be interested in seeing more of the uses.... --------- On the conditional assignment topic, since no other ex-Multifloids spoke up yet, I'll just mention that the Trace family had multiple condition codes and could do a "select" operation between two values based on a condition code bit. The new machine we were working on upon our demise.., , had both selects and conditional assignments. In a VLIW it isn't very hard to implement this critters. In a typical scoreboarded machine you might not like the select instructions as they are 4-address instructions. In addition, the floating point data path might not have the ability to transport either source field, etc, etc. pkr -- Paul K. Rodman Advanced Systems Division rodman@sgi.com Silicon Graphics, Inc. KA1ZA Mountain View, Ca.