Re: Can an intrinsic function be replaced in FTN95


I assume you must be talking about the issue of signed zeros.

If you have a 100,000-line program in f90, you should probably
be using modules anyway.

Thanks Richard and Michael. Making it an external function worked well. That
was the problem I was having. The original authors were using sign to test for
negative zeros (why, I do not know.)

I was tasked with getting it to work under FTN95 and it was blowing up on the
same data that worked under FTN90. I traced it to the sign function behaving
differently under the newer compiler.

I've been a coder for years, but hadn't played with FORTRAN again until about 6
months ago, and was rusty about how to get around this. Your help was
appreciated!

Thanks, Tom

.

Relevant Pages

  • Honor Roll of Donors; Douglas Andrew Kukucka
    ... John Anthony Bielec ... Thomas James Drach ... Michael G. Nishizaki ... Richard J. Piluk ...
    (sci.engr.civil)
  • Re: Interpol "configured to reject your mail" re Paul Gieschen (Heropoules/Vensel)
    ... Richard, Chuck White and Kevin Smith are responsible for the most ... "There's no way he (God) could have done this. ... Tom ordered Michael Doberson to be patient with me in the Microbiology ...
    (sci.geo.satellite-nav)
  • Halloween Swingthing Pass Winners List!
    ... Dustin Anderson, Aaron Batista, Martin Brown, Frank Carricato, Josh Clark, ... Alfred Lee, Rodger Magill, Richard Maruyama, Ben McHenry, Joe Pawloski, Greg ... Dearborn, Jordan Frisbee, Gary Jobst, Michael Kiehm, Richard Kear, Michael ...
    (rec.arts.dance)
  • Re: Unfall
    ... Michael Buege schrub: ... > Richard, du erklaerst gerade einem Schneemann, was ein ... Das Bewusstsein bestimmt das Sein. ... Next by Date: ...
    (de.rec.fahrrad)
  • Re: Q about F90 Array Manipulation
    ... Richard: ... I am in the process of shifting from F77 to F90, and my mind is still ...
    (comp.lang.fortran)