Floating Point Verification

Moderator, John S. Quarterman std-unix at longway.TIC.COM
Fri Jul 15 07:23:45 AEST 1988


From: uunet!gatech.edu!ut-sally!ut-emx!chpf127 (J. Eaton)

In article <208 at longway.TIC.COM>
> From: uunet!tahoe.unr.edu!bryson (Derry Bryson)
> 
> I am searching for a standard to check floating point processors and 
> emulators against.  Basically what I would like to find is a program 
> that performs several operations and evaluates the accuracy of the answers
> based upon some kind of accepted standard.  If anyone has any information
> or advice I would greatly appreciate hearing from you.
...
> Volume-Number: Volume 14, Number 28

  There is a program called PARANOIA available from netlib at anl-mcs.arpa
  (the online software library maintained at Argonne National Laboratory)
  which does what you want.  There are five versions I know about, one
  each in C, Pascal, and Basic, and two in Fortran (double/single).
  This program tries to test the floating point processor and/or library of
  floating point subroutines.  If the machine you use has one library
  which all languages share it shouldn't matter what version you use
  (but running all of them might prove to be interesting...)

  To get a copy of one of the versions, send a message of the form:

    send <item> from paranoia

  <item> is one of:  paranoia.b, paranoia.c, paranoia.p, spara.f, or dpara.f

  your message can include more than one request, but only one request
  per line.  You might also want to get the general netlib index, by sending
  the message:

    send index


  J. Eaton
  UT Department of Chemical Engineering

  Not really doing anything with chemicals.

Volume-Number: Volume 14, Number 31



More information about the Comp.std.unix mailing list