Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!dali.cs.montana.edu!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!wuarchive!emory!ducvax.auburn.edu!noc.sura.net!europa.asd.contel.com!gatech!usenet.ins.cwru.edu!tut.cis.ohio-state.edu!VAX1.CC.UAKRON.EDU!r2sms
From: r2sms@VAX1.CC.UAKRON.EDU (Sujat M Sukthankar)
Newsgroups: comp.lang.c
Subject: C FLOATING POINT BUG - OR IS IT ?
Message-ID: <1991May24.175749.4263@ducvax.auburn.edu>
Date: 24 May 91 22:57:48 GMT
Lines: 25


                                                  
                  - A CHALLENGE
Message-ID: <1378@VAX1.CC.UAKRON.EDU>
Date: 24 May 91 14:31:16 GMT
Reply-To: r2sms@VAX1.CC.UAKRON.EDU (Sujat M Sukthankar)
Distribution: usa
Organization: U of Akron, Ohio
Lines: 15


      The following is a major/minor bug which maybe/may not be trivial.

Try assigning a "float" variable to a "double" variable - results in a
loss of accuracy with the magnitude being greater for larger numbers. A
"promotion" should result in better accuracy and not viceversa. A simple
rogram was tested on the Microsoft Ver 6 compiler and also the watcom
386 compiler. The problem started off as trivial but still hasn't been
solved.             

                     
      
       See if you any of you guys out there can do it.

n a better accuracy