Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!mailrus!usenet.ins.cwru.edu!cwjcc!ncoast!util20!sir-alan!mikes
From: mikes@sir-alan.UUCP (Mike Squires)
Newsgroups: comp.sys.next
Subject: Numerical errors in NeXT FORTRAN
Keywords: FORTRAN errors
Message-ID: <197@sir-alan.UUCP>
Date: 1 Feb 90 04:39:35 GMT
Organization: Meadville Public XENIX, Meadville, PA
Lines: 16

The FORTRAN compiler available for the NeXT (Absoft?) seems to produce
slightly different numerical results as compared to the same code
run on a Sun 4/110 using Sun FORTRAN or a 386 XENIX system running
LPI FORTRAN 77.  The errors are large enough to make the use of the
code problematical (the software uses non-linear regression and
interates to convergence, but the errors show up in the first
iteration).

Has anyone else seen this?

Michael L. Squires           uucp: ..!{convex,pitt,util20,uunet}!sir-alan!mikes
752 Chestnut Street                       ..!cwjcc!ncoast!util20!sir-alan!mikes
Meadville, PA 16335                      BITNET: mikes%sir-alan@pitt.UUCP (VAX)
Voice:  814 337 5528                          MIKES AT SIR-ALAN!PITT.UUCP (IBM)
Data:   814 337 0348             Internet: sir-alan!mikes@uunet.uu.net         
login of "ubbs" for BBS                    mikes@ncoast.org