| Re: problem converting FORTRAN to IDL [message #4048 is a reply to message #3962] |
Sat, 15 April 1995 00:00  |
rivers
Messages: 228
Registered: March 1991
|
Senior Member |
|
|
In article <3mlnpd$qnd@reznor.larc.nasa.gov>, zawodny@arbd0.larc.nasa.gov (Joseph M Zawodny) writes:
> In article <D6zHn2.LyF@ireq.hydro.qc.ca> brooker@toka.ireq-ccfm.hydro.qc.ca writes:
>> This is an observation I have just made about IDL.
>>
>> When you compile a FORTRAN program, you can specify G_floating
>> implementations of REAL*8. This extends the range of numbers to +-0.56D308.
>> (For default D_floating, the maximum number allowed is 0.29D38.)
>>
>> On the other hand, IDL has no option for the larger G_floating numbers. This
>> makes for problems when you convert a "G_floating REAL*8 " FORTRAN program to
>> IDL.
>>
>> Peter Brooker
>
> Gee, maybe I do not understand your problem, but I did this quick test.
>
> IDL> a=.5d308
> IDL> print,a
> 5.0000000e+307
>
> Therefor you should be able to use the IDL DOUBLE to implement FORTRAN
> G_floating calculations.
I belive the original post was probably referring to a DEC Alpha machine. On
the Alpha one can compile double precision code to be D_FLOAT, G_FLOAT or
IEEE_FLOAT. The default for the DEC C compiler is G_FLOAT. IDL is clearly
compiled with D_FLOAT, presumably so that the Alpha version is compatible with
old VAX binary files, where D_FLOAT is the default. It would be possible and
perhaps nice if RSI would provide 3 different versions (D_FLOAT, G_FLOAT and
IEEE_FLOAT) of IDL for the Alpha platform. If you use CALL_EXTERNAL on the
Alpha you need to make sure any routines you call are compiled with the same
floating point format used for IDL.
____________________________________________________________
Mark Rivers (312) 702-2279 (office)
CARS (312) 702-9951 (secretary)
Univ. of Chicago (312) 702-5454 (FAX)
5640 S. Ellis Ave. (708) 922-0499 (home)
Chicago, IL 60637 rivers@cars3.uchicago.edu (Internet)
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
