| IDL for Windows 3.0: Any more info? [message #257] |
Wed, 18 March 1992 13:48  |
howp
Messages: 3
Registered: March 1992
|
Junior Member |
|
|
Hello, IDL users:
I have heard that IDL for the PC is coming out in August/92. This
is what a fellow graduate student found out after talking to the IDL
company (RSI). Apparently they had a beta version of IDL for DOS but
it was so slow that they decided to abandon the program and the next
implementation of IDL will be for Microsoft Windows 3.0.
My questions are:
1) Has anybody heard anything more about this ( or is this vapourware )?
2) Has anybody used the beta version of IDL for Windows ( if it exists yet )
and where can I get a copy?
Thanks for your time.
Peter How ( Graduate Student )
Institute of Space and Atmospheric Studies,
University of Saskatchewan,
Saskatoon, SK
CANADA
S7N 0W0
I can be reached at
(306) 966-6452
IN%"HOW@SKISAS.USASK.CA"
IN%"HOWP@SKYFOX.USASK.CA"
|
|
|
|
|
|
| Re: idl [message #291 is a reply to message #257] |
Tue, 25 February 1992 09:41  |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
In article <1992Feb25.150743.20626@cs.tulane.edu>, terrie@merl.noarl.navy.mil
(Greg Terrie) writes...
>
> I am trying to call a fortran subroutine from IDL using the call_external
> function. I am running on a SUN 4/390 f77 version 1.3.1, SUNOS 4.1.1. I have
> followed the instructions for creating a shareable library as described in the
> IDL manual. However when I issue the command from the IDL prompt:
>
> iret=call_external('fcode.so','_test_')
>
> I get the error message:
>
> ld.so: call to undefined procedure _s_wsle from 0xf????????
>
I think the problem is that you are trying to call the Fortran routine
directly. IDL on the Sun doesn't want to do this (although it can in VMS).
What I do (and I think this came from RSI originally) is to create a C-program
to sit between IDL and the FORTRAN program. For instance,
#include <stdio.h>
void interpol_c(argc, argv)
int argc; /* The number of arguments */
void *argv[]; /* The arguments */
{
float *x_in, *y_in, *x_out, *y_out;
long *n_in, *n_out;
/* Convert the IDL input parameters into FORTRAN parameters. */
x_in = (float *) argv[0];
y_in = (float *) argv[1];
n_in = (long *) argv[2];
x_out = (float *) argv[3];
y_out = (float *) argv[4];
n_out = (long *) argv[5];
/* Call the FORTRAN routine INTERPOL. */
interpol_(x_in,y_in,n_in,x_out,y_out,n_out);
return;
}
The Makefile for this would include the commands
cc -pic -fsingle -c interpol_c.c
f77 -fast -pic -c interpol.f
ld -o external.so -assert pure-text interpol_c.o interpol.o
Then call_external attaches to "_interpol_c" rather than "_interpol_".
Bill Thompson
|
|
|
|
| Re: idl [message #292 is a reply to message #257] |
Tue, 25 February 1992 09:39 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
In article <1992Feb25.150743.20626@cs.tulane.edu>, terrie@merl.noarl.navy.mil
(Greg Terrie) writes...
>
> I am trying to call a fortran subroutine from IDL using the call_external
> function. I am running on a SUN 4/390 f77 version 1.3.1, SUNOS 4.1.1. I have
> followed the instructions for creating a shareable library as described in the
> IDL manual. However when I issue the command from the IDL prompt:
>
> iret=call_external('fcode.so','_test_')
>
> I get the error message:
>
> ld.so: call to undefined procedure _s_wsle from 0xf????????
>
I think the problem is that you are trying to call the Fortran routine
directly. IDL on the Sun doesn't want to do this (although it can in VMS).
What I do (and I think this came from RSI originally) is to create a C-program
to sit between IDL and the FORTRAN program. For instance,
#include <stdio.h>
void interpol_c(argc, argv)
int argc; /* The number of arguments */
void *argv[]; /* The arguments */
{
float *x_in, *y_in, *x_out, *y_out;
long *n_in, *n_out;
/* Convert the IDL input parameters into FORTRAN parameters. */
x_in = (float *) argv[0];
y_in = (float *) argv[1];
n_in = (long *) argv[2];
x_out = (float *) argv[3];
y_out = (float *) argv[4];
n_out = (long *) argv[5];
/* Call the FORTRAN routine INTERPOL. */
interpol_(x_in,y_in,n_in,x_out,y_out,n_out);
return;
}
The Makefile for this would include the commands
cc -pic -fsingle -c interpol_c.c
f77 -fast -pic -c interpol.f
ld -o external.so -assert pure-text interpol_c.o interpol.o
Then call_external attaches to "_interpol_c" rather than "_interpol_".
Bill Thompson
|
|
|
|
| Re: idl [message #293 is a reply to message #257] |
Tue, 25 February 1992 09:14 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
In article <1992Feb25.150743.20626@cs.tulane.edu>, terrie@merl.noarl.navy.mil
(Greg Terrie) writes...
>
> I am trying to call a fortran subroutine from IDL using the call_external
> function. I am running on a SUN 4/390 f77 version 1.3.1, SUNOS 4.1.1. I have
> followed the instructions for creating a shareable library as described in the
> IDL manual. However when I issue the command from the IDL prompt:
>
> iret=call_external('fcode.so','_test_')
>
> I get the error message:
>
> ld.so: call to undefined procedure _s_wsle from 0xf????????
>
I think the problem is that you are trying to call the Fortran routine
directly. IDL on the Sun doesn't want to do this (although it can in VMS).
What I do (and I think this came from RSI originally) is to create a C-program
to sit between IDL and the FORTRAN program. For instance,
#include <stdio.h>
void interpol_c(argc, argv)
int argc; /* The number of arguments */
void *argv[]; /* The arguments */
{
float *x_in, *y_in, *x_out, *y_out;
long *n_in, *n_out;
/* Convert the IDL input parameters into FORTRAN parameters. */
x_in = (float *) argv[0];
y_in = (float *) argv[1];
n_in = (long *) argv[2];
x_out = (float *) argv[3];
y_out = (float *) argv[4];
n_out = (long *) argv[5];
/* Call the FORTRAN routine INTERPOL. */
interpol_(x_in,y_in,n_in,x_out,y_out,n_out);
return;
}
The Makefile for this would include the commands
cc -pic -fsingle -c interpol_c.c
f77 -fast -pic -c interpol.f
ld -o external.so -assert pure-text interpol_c.o interpol.o
Then call_external attaches to "_interpol_c" rather than "_interpol_".
Bill Thompson
|
|
|
|
|
|
| Re: IDL for Windows 3.0: Any more info? [message #399 is a reply to message #257] |
Fri, 20 March 1992 07:26 |
bradleyt
Messages: 4
Registered: March 1992
|
Junior Member |
|
|
In article <1992Mar20.060055.21428@ucsu.Colorado.EDU> idl@boulder.colorado.edu
(Keith Crosley) in the guise of bradleyt@ucsu.Colorado.EDU (God) writes:
>
> Obviously, the previous poster is a bit misinformed. Since I have
> worked with both versions of IDL for the PC, I can attest to the fact
> that the Windows version of IDL *is* faster. Part of the reason is that
> the Windows port is simply better done than the previous PC port...
> better programming, etc. Also, since the original PC version had to have
> built-in drivers for many video cards that were not neccessarily optimized,
> many users will note a speed up when they use their manufacturer-supplied
> Windows graphics drivers.
First, let me explain a few things to those readers who may be
a little confused at this point. No, I am not replying to my
own post. A friend of mine named Keith Crosley (see below)
who works at RSI, the manufacturer of IDL, has been using one
of my computer accounts to read mail and news. However, as part
of the University's agreement for getting a computer account,
my login may not be used for any commercial projects. Let me
apologize for giving anybody the impression that the account
bradleyt@ucsu.colorado.edu has anything to do with Research
Systems Inc., a for-profit company. Keith may be contacted at
his address below, but not by replying to the previously posted
message.
> Anyway, Todd should know this information already as he has seen both
> versions of IDL for the PC since he is a close friend of mine and
> partner in industrial music crime.
For those of you unacquainted with the warped minds of RSI's
"Technical Communications" department, this is Keith's idea
of a joke. Actually, I have seen neither the DOS version nor
the Windows version of IDL. However, I would be more than happy
to offer to be part of RSI's BETA program for IDL for Windows.
Does anyone know where I can apply for this program?
> Keith R Crosley
> Director of Technical Communication
> Research Systems, Inc. / 777 29th Street / Boulder, CO 80303 / (303)786-9900
> e-mail: idl@boulder.colorado.edu
If you are trying to contact Mr. Crosley, be sure to send mail to
the above address, not to bradleyt@ucsu.colorado.edu.
--
Todd "Slan" Bradley -- Postmodern Reaganomics Renaissance Man, The
Exception to Every Rule, An Example To All Men, The Biggest Jerk In
The World, Supreme Ruler Of The Galaxy, AND Captain of The Flying
Bitheads Ultimate Frisbee Team. (303) 443-6317 or 492-5826
|
|
|
|
| Re: IDL for Windows 3.0: Any more info? [message #400 is a reply to message #257] |
Thu, 19 March 1992 22:00 |
bradleyt
Messages: 4
Registered: March 1992
|
Junior Member |
|
|
In article <1992Mar20.034534.10318@ucsu.Colorado.EDU> bradleyt@spot.Colorado.EDU (Todd "Slan" Bradley) writes:
> In article <18MAR92.21482584@skyfox.usask.ca> howp@skyfox.usask.ca writes:
>
>> I have heard that IDL for the PC is coming out in August/92. This
>> is what a fellow graduate student found out after talking to the IDL
>> company (RSI). Apparently they had a beta version of IDL for DOS but
>> it was so slow that they decided to abandon the program and the next
>> implementation of IDL will be for Microsoft Windows 3.0.
>
>
> So you're saying it is too slow to run under DOS, so they
> made it run under Windows 3.0? Talk about a diverging
> solution...
As if PVI even HAS a scientific visualization solution for the PC!
>
> Todd.
> ps. I would, of course, love to hear ANY attempt to explain
> how a Windows port is going to make any program run faster.
>
> --
> Todd "Slan" Bradley -- Postmodern Reaganomics Renaissance Man, The
> Exception to Every Rule, An Example To All Men, The Biggest Jerk In
> The World, Supreme Ruler Of The Galaxy, AND Captain of The Flying
> Bitheads Ultimate Frisbee Team. (303) 443-6317 or 492-5826
Well, Todd, I suppose you'll forgive me for using your account to
post this followup... I just can't resist!
Obviously, the previous poster is a bit misinformed. Since I have
worked with both versions of IDL for the PC, I can attest to the fact
that the Windows version of IDL *is* faster. Part of the reason is that
the Windows port is simply better done than the previous PC port...
better programming, etc. Also, since the original PC version had to have
built-in drivers for many video cards that were not neccessarily optimized,
many users will note a speed up when they use their manufacturer-supplied
Windows graphics drivers.
I hope that explains a bit better why the Windows version is, in many
ways, faster than the old version. The Windows device is also much
more like the "X" device. You can have multiple windows, cursor
control, etc. In the DOS version, you were stuck with alternating
between a text-only and a graphics-only screen.
Anyway, Todd should know this information already as he has seen both
versions of IDL for the PC since he is a close friend of mine and
partner in industrial music crime.
Also, to correct the previous post: The Windows version of IDL has been
in ALPHA testing. We got some good feedback and right now, just about
everything seems to work well. We will be announcing availability of the
Windows BETA release (which will include a free upgrade to the final
release version) in the April IDL newsletter. I'm working on the Windows
version installation and release notes now. They'll be done before I
go on my well-earned vacation in early April.
Say "hi" to everyone at PVI for me, Todd!!
Keith R Crosley
Director of Technical Communication
Research Systems, Inc. / 777 29th Street / Boulder, CO 80303 / (303)786-9900
e-mail: idl@boulder.colorado.edu
|
|
|
|
| Re: IDL for Windows 3.0: Any more info? [message #401 is a reply to message #257] |
Thu, 19 March 1992 19:45 |
bradleyt
Messages: 4
Registered: March 1992
|
Junior Member |
|
|
In article <18MAR92.21482584@skyfox.usask.ca> howp@skyfox.usask.ca writes:
> I have heard that IDL for the PC is coming out in August/92. This
> is what a fellow graduate student found out after talking to the IDL
> company (RSI). Apparently they had a beta version of IDL for DOS but
> it was so slow that they decided to abandon the program and the next
> implementation of IDL will be for Microsoft Windows 3.0.
So you're saying it is too slow to run under DOS, so they
made it run under Windows 3.0? Talk about a diverging
solution...
Todd.
ps. I would, of course, love to hear ANY attempt to explain
how a Windows port is going to make any program run faster.
--
Todd "Slan" Bradley -- Postmodern Reaganomics Renaissance Man, The
Exception to Every Rule, An Example To All Men, The Biggest Jerk In
The World, Supreme Ruler Of The Galaxy, AND Captain of The Flying
Bitheads Ultimate Frisbee Team. (303) 443-6317 or 492-5826
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
