| NaN or 0.0 [message #8969] |
Tue, 20 May 1997 00:00  |
Christophe Morisset
Messages: 2
Registered: March 1997
|
Junior Member |
|
|
Hello,
I'm "transporting" an IDL code from a DecAlpha to a PC (i.e. Unix =>
Linux)
And I meet problems with the NaN under Linux:
Under Unix:
IDL> print,exp(alog(0))
% Program caused arithmetic error: Floating overflow
% Detected at $MAIN$
% Program caused arithmetic error: Floating underflow
% Detected at $MAIN$
0.00000
Under Linux:
IDL> print,exp(alog(0))
-NaN
% Program caused arithmetic error: Floating illegal operand
In both case, the program say I'm doing something wrong,
but with the Unix version, I can continue to work...
Is it possible to say the Linux-IDL not to use this
&*%$%# NaN absorbing element?
It seems to me it as something to do with something
called IEEE, but I don't understand what it is ;-)
Tahnks a lot if you have any idea,
--
Christophe Morisset
Observatoire de Paris-Meudon
Tel: (+33) 1-45-07-75-66
Fax: (+33) 1-45-07-74-69
mailto:morisset@daec.obspm.fr
http://daec.obspm.fr/~morisset
|
|
|
|
| Re: NaN or 0.0 [message #9047 is a reply to message #8969] |
Wed, 21 May 1997 00:00  |
marq
Messages: 12
Registered: February 1996
|
Junior Member |
|
|
Hi Christophe,
you wrote:
> Under Linux:
> IDL> print,exp(alog(0))
> -NaN
> % Program caused arithmetic error: Floating illegal operand
>
> In both case, the program say I'm doing something wrong,
> but with the Unix version, I can continue to work...
>
> Is it possible to say the Linux-IDL not to use this
> &*%$%# NaN absorbing element?
> It seems to me it as something to do with something
> called IEEE, but I don't understand what it is ;-)
...it's just a standard how to handle floating point exceptions. On
IEEE machines, invalid math operations (like your one) result in
an NaN (i.e., Not a Number) or something like 'Infinity', plus a signal
delivered to the program, idl in this case. It is then up to the program
how to react to the floating point exception - may that be halting (like
VMS always does) or just reporting the occurence of the exception, but
continuing as if nothing had happened.
With my linux version of IDL (which is 5.0 beta 6), a system variable
'!EXCEPT' exists which controls when IDL issues its warnig: never
(!EXCEPT = 0), _after_ a function or procedure has finished
(!EXCEPT = 1, which seems to be the default), or immediately after
the program line that caused the exception (!EXCEPT = 2).
In your case:
IDL> print, !except
1
IDL> print, exp(alog(0))
NaN
% Program caused arithmetic error: Floating illegal operand
IDL> !except = 2
IDL> print, exp(alog(0))
NaN
% Program caused arithmetic error: Floating illegal operand
% Detected at $MAIN$
Hope this helps a bit...
Regards,
Chris.
------------------------------------------------------------ ---------------
Christian Marquardt
Meteorologisches Institut der | tel.: (+49) 30-838-71170
Freien Universitaet Berlin | fax.: (+49) 30-838-71167
Carl-Heinrich-Becker-Weg 6-10 | email: marq@strat01.met.fu-berlin.de
D-12165 Berlin |
------------------------------------------------------------ ---------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
