vis5d-users Mailing List for Vis5d+

> Hi - sorry to post a daft question but I've looked around and cant
> find an answer. What do you use for X/OpenGL in MacOS X? I know that
> aqua supports it but "configure"  in Vis5d.

You can download XFree86 for MacOS X from the XonX project:

	http://mrcla.com/XonX/

(Actually, the official XFree86 4.1 now supports MacOS, but you need to
get the latest developer snapshot, XDarwin 1.0a1, in order to have OpenGL
support.  (At one point, you had to install Mesa too, just to get the
headers, but I think they may have fixed that.)

Steven

Hi - sorry to post a daft question but I've looked around and cant 
find an answer. What do you use for X/OpenGL in MacOS X? I know that 
aqua supports it but "configure"  in Vis5d.

Steven :- the one you posted below is looking for Mesa - Mesa wont 
compile without a "GLUT" - I'm confused. Thanks if you have time to 
reply to this.

Vic.

PS.
I'm really looking forward to using Vis5D on my mac - it would be 
even nicer if apple would do something about Java3D so that I can 
develop VisAD solutions there too....

>On Wed, 25 Jul 2001, Edward (Ted) Mansell wrote:
>>  I've managed to compile vis5d (versions 4.3 and 5.1) under Mac OS X
>>  using Apple's version of gcc and with XFree86 installed.  (I mainly
>>  used the freebsd-x options except for -DLITTLE).  It does fine with
>>  isosurfaces and colored slices, but crashes on contour slices
>>  ("segmentation fault" in contour.c -- gdb says something about an
>>  address problem) and freezes when trying to clone a variable.
>
>I know that a number of bugs have been fixed since version 5.1... 
>
>I also know (because I use MacOS X myself at work) that the CVS version of
>Vis5d+ (vis5d.sourceforge.net) recently compiled "out of the box" on MacOS
>X (assuming you have the latest XFree86 from XonX installed), and contour
>slices work just fine.
>
>Unfortunately, Vis5d+ 1.1 doesn't quite compile out of the box; a couple
>of patches were needed because I forgot to test that release on a machine
>(like MacOS X) lacking a Fortran compiler, which is why I point to the CVS
>version above.  (I'll try to get a Vis5d+ 1.1.1 out the door in a short
>time to fix this problem.)
>
>Steven
>
>PS. For the impatient, I've posted a snapshot of the current 1.1.1 alpha
>at: http://ab-initio.mit.edu/download/vis5d+-1.1.1a1.tar.gz (I haven't
>tested this...just did a quick 'make dist'...let me know if it works).


-- 
_____________________________________________________
Vic Cornell.

IT Development Manager.
Information Technology Group,
Room 256/29, Southampton Oceanography Centre,
Empress Dock,  European Way,
SOUTHAMPTON, SO14 3ZH, UK.
Email: vc...@ma...	http://www.soc.soton.ac.uk/
Tel: 02380 596193		Fax: 02380 596195
_____________________________________________________

Hi Kevin,

> Thanks to those that helped me out earlier with compiling the maketopo.c.  I
> knew there was something simple and stupid, but didn't realize that it was that
> simple.
> 
> Of course, getting that compiled didn't help out too much, since I have not been
> able to find any information as to what the format for the input file should be,
> nor a reference on how to use the command. I am using, for now, a simple
> 1:250,000 DEM file from the USGS.  I have read from a website at the Univ or
> Montana that the header for this has to be chopped out.  What is Vis5d looking
> for when importing a file??  Would it possibly be easier to import a simply
> x,y,z for the topo, or is the DEM the way to go??
> p.s. if anyone out there has some notes on vis5d, would it be possible to get a
> glance at them.  I have searched throughout the web for answers to questions,
> but no sites really help with the import of files.  Thanks....

maketopo.c does not define an input format. Rather, you need
to add your own code to maketopo.c to read whatever input file
you have, and put the data into arrays acording to the comments
at the top of maketopo.c.

For notes about Vis5D, check out the Vis5D README file, which is
linked from the Vis5D web page and included with the source
distribution.

Good luck,
Bill
----------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hi...@fa...  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html

Thanks to those that helped me out earlier with compiling the maketopo.c.  I
knew there was something simple and stupid, but didn't realize that it was that
simple.

Of course, getting that compiled didn't help out too much, since I have not been
able to find any information as to what the format for the input file should be,
nor a reference on how to use the command. I am using, for now, a simple
1:250,000 DEM file from the USGS.  I have read from a website at the Univ or
Montana that the header for this has to be chopped out.  What is Vis5d looking
for when importing a file??  Would it possibly be easier to import a simply
x,y,z for the topo, or is the DEM the way to go??
p.s. if anyone out there has some notes on vis5d, would it be possible to get a
glance at them.  I have searched throughout the web for answers to questions,
but no sites really help with the import of files.  Thanks....

--
Kevin Slover, Meteorologist/Oceanographer
Tropical Prediction Center/National Hurricane Center
11691 SW 17th Street Miami FL 33165

Work: (305) 229-4456 Fax : (305) 553-1264
e-mail:  kjs...@nh...  kev...@no...
Web address:  http://www.nhc.noaa.gov

Hi Rashaad,

I'm not sure if anyone has responded to you yet, but there was an error on
the Cave5D website concerning VIS5D_MEM_SIZE.  The actual variable is
VIS5D_MEMSIZE and similarly C5D_MEMSIZE.  Right now I currently have the
code for Cave5D and have made many changes to it. I'm planning to release
version 2.0 shortly, but have no date yet.  I go into a little bit more
detail about it on my web site (http://www-unix.mcs.anl.gov/~voelz/).

-Sheri



Sheri Voelz			
Mathematics & Computer Science
Argonne National Laboratory	                
9700 South Cass Ave, Argonne, Illinois 60439   
Tel: 630-252-7526


On Wed, 18 Jul 2001, Rashaad Jones wrote:

> Is anyone using Cave5d to visualize vis5d data?  If so, how are you
> expanding the memory on Cave5d?  I have Cave5d version 1.4 (only the
> executable) and I cannot change the default's memory.  Cave5d version
> 1.4 has 2 memory parameters that can be set to define the memory
> (C5D_MEM_SIZE and VIS5D_MEM_SIZE), however cave5d ignores these
> paramters when I set them in my config file.
> Also, does anyone know where I can obtain the source code for cave5d?
> 
> Rashaad Jones
> Research Assistant
> Applied Research Laboratory
> Pennsylvania State University
> 
> 
> 
> 
> 
> 

If you're using Vis5d+, you can just do 'make maketopo'.  

I am fairly new to the Vis5d world, and am having a major issue with
compiling the maketopo.c program.  I have been searching around looking
for help, and found that the makefile (maketopo.m) is needed for those,
though it was not included (at least from what I have seen) in the
download and installation.  Is there some outside source that I would
have to go to download this file??  Or am I making a simple error and
just not seeing it??

--
Kevin Slover, Meteorologist/Oceanographer
Tropical Prediction Center/National Hurricane Center
11691 SW 17th Street Miami FL 33165

Work: (305) 229-4456 Fax : (305) 553-1264
e-mail:  kjs...@nh...  kev...@no...
Web address:  http://www.nhc.noaa.gov

On Wed, 25 Jul 2001, Edward (Ted) Mansell wrote:
> I've managed to compile vis5d (versions 4.3 and 5.1) under Mac OS X
> using Apple's version of gcc and with XFree86 installed.  (I mainly
> used the freebsd-x options except for -DLITTLE).  It does fine with
> isosurfaces and colored slices, but crashes on contour slices
> ("segmentation fault" in contour.c -- gdb says something about an
> address problem) and freezes when trying to clone a variable.

I know that a number of bugs have been fixed since version 5.1...  

I also know (because I use MacOS X myself at work) that the CVS version of
Vis5d+ (vis5d.sourceforge.net) recently compiled "out of the box" on MacOS
X (assuming you have the latest XFree86 from XonX installed), and contour
slices work just fine.

Unfortunately, Vis5d+ 1.1 doesn't quite compile out of the box; a couple
of patches were needed because I forgot to test that release on a machine
(like MacOS X) lacking a Fortran compiler, which is why I point to the CVS
version above.  (I'll try to get a Vis5d+ 1.1.1 out the door in a short
time to fix this problem.)

Steven

PS. For the impatient, I've posted a snapshot of the current 1.1.1 alpha
at: http://ab-initio.mit.edu/download/vis5d+-1.1.1a1.tar.gz (I haven't
tested this...just did a quick 'make dist'...let me know if it works).

Hello,

Sorry if this has been discussed already, but I'm new to the mailing list.

I've managed to compile vis5d (versions 4.3 and 5.1) under Mac OS X using Apple's version of gcc and with XFree86 installed.  (I mainly used the freebsd-x options except for  -DLITTLE).  It does fine with isosurfaces and colored slices, but crashes on contour slices ("segmentation fault" in contour.c -- gdb says something about an address problem) and freezes when trying to clone a variable.  

If there is any interest or suggestions, just let me know.

-- Ted

"Rashaad E. Jones" wrote:

> I don't understand how vis5d uses its memory.  I have v5d files that are
> between 20 mbs and 50 mbs in size, however, I need to specify vis5d to
> use at least 512 mbs of memory (and sometimes that doesn't work).  Can
> anyone explain why?
>
> Rashaad Jones

Rashaad,

If you have 512 MB of memory you should probably set mbs=0 and just use the
system malloc.   The amount of memory required by vis5d depends on the
complexity of the data and the graphics you are trying to produce as well
as the size of the data files.  Also I believe that the vis5d+ distribution
fixes some memory problems, you might try downloading it.

Jim

I don't understand how vis5d uses its memory.  I have v5d files that are
between 20 mbs and 50 mbs in size, however, I need to specify vis5d to
use at least 512 mbs of memory (and sometimes that doesn't work).  Can
anyone explain why?

Rashaad Jones

On Tue, 24 Jul 2001, Ryan K. Decker wrote:
> I am tying to customize vis5d to be able to handle more times than the 
> default 400.  I have changed MAXTIMES to 800 in both the v5d.h and v5df.h 
> files and have recompiled the makefile (make linux-x) like the instructions 
> suggest.  However, when I run my program to create the v5d file with more 
> than 400 frames I get this error message:
> 
> Too many timesteps: 421  (Maximum is 400)
> Error in v5dCreateSimpleFile: unable to create 14hRun01.v5d

The only explanation that I can see is that you didn't recompile
everything.  First, I'm not sure that the dependencies are correct in the
Vis5d 5.2 makefile, so you may have to manually remove *.o before
recompiling.  Of course, make sure you recompile your own program, too.

Steven

Everyone,

I am tying to customize vis5d to be able to handle more times than the 
default 400.  I have changed MAXTIMES to 800 in both the v5d.h and v5df.h 
files and have recompiled the makefile (make linux-x) like the instructions 
suggest.  However, when I run my program to create the v5d file with more 
than 400 frames I get this error message:

Too many timesteps: 421  (Maximum is 400)
Error in v5dCreateSimpleFile: unable to create 14hRun01.v5d

Does anyone have any insight into my problem? 
-- 
Ryan


*********************************************************************************
Ryan K. Decker                phone: (865)-576-2326
Research Associate         fax: (865)-567-1327
UAH/ESSC                      email: de...@at...
456 S. Illinois Ave.            home page: http://vortex.nsstc.uah.edu/~decker
PO Box 2456
Oak Ridge, TN 37831
********************************************************************************

I'm running vis5d version 5.2 and I'm having trouble understanding how
vis5d displays its isosurfaces.  I created a isosurface for a particular
isovalue, and vis5d generated an isosurface, however this isosurface
covers too much area on the x, y plane.  Has anyone else experienced
this?

Rashaad

Eileen,

set it to 0 - this tells vis5d not to use its own internal memory managment
and to use malloc free instead.






Eileen Rotthoff wrote:

> Hi,
>
> Does anyone know what is the maximum memory size that can be specified for
> the Vis5D executable?  It seems to increase memory usage for us until we
> specify a size over 512.  We have plenty of memory on the Onyx2 (8 Gbytes)
> so it seems like it might be a limitation in Vis5D.
>
> Thanks,
> Eileen Rotthoff
>
> Penn State University
> Applied Research Lab

Hi,

Does anyone know what is the maximum memory size that can be specified for 
the Vis5D executable?  It seems to increase memory usage for us until we 
specify a size over 512.  We have plenty of memory on the Onyx2 (8 Gbytes) 
so it seems like it might be a limitation in Vis5D.

Thanks,
Eileen Rotthoff

Penn State University
Applied Research Lab

Dear Colleagues,

The problem that I have is that my  'vis5d' display package cannot
display the country
map overlay for our East African domain , it only displays the
continental boundaries
and lakes. Please , kindly assist me in solving this problem.

Regards,

John.

Hi,


Several of the calls in work.c to

extract_hslicePRIME

use ctx->LowLev[var] while others use dtx->LowLev and it looks to me
like  the dtx->LowLev calls are wrong, but I'm not sure yet.

If I have a 3d var with LowLev > 0 I shouldn't expect to see data
contours in at the lowest level (below the valid range of that variable)
should I?

Can anyone clarify this usage?

Thanks,

Jim

On Wed, 18 Jul 2001, Fritz Kristina Contractor HQ AFWA/SAIC wrote:
> I have a non-technical question concerning the recent fracas over the GIF
> format using LZW compression.  Does anyone know how vis5d resolved the
> issue?  I'm asking because we use vis5d in our organization and I'm
> responsible for making sure we don't get any surprises.

Vis5D doesn't create GIF images.  (It has some support for them on SGI
systems, but there it uses the external "togif" program.)

(We really should just add PNG support.)

Steven

I have a non-technical question concerning the recent fracas over the GIF
format using LZW compression.  Does anyone know how vis5d resolved the
issue?  I'm asking because we use vis5d in our organization and I'm
responsible for making sure we don't get any surprises.

Kristina L. Fritz
Senior Functional Analyst, SAIC
HQ AFWA/XPSI Offutt AFB, NE
* DSN: 271-0847
* COMM: (402) 294-0847
* kri...@af...

Hi -
A resource for map outlines is this link:
http://rimmer.ngdc.noaa.gov/coast/getcoast.html

You can request country and other political boundaries,
rivers, and coastlines.   I can't help with getting these into
the form you'll need to display them with Vis5D, but it's
a start.

Ansley Manke

john mungai wrote:

> Hi
>
> > Am currently running vis5d to display some weather model output over
> Africa.
> > The main problem that I have is that I can only overlay the continetal
>
> > boundary
> > and not the country's natioanl boundary map. Please assist me on how I
> can
> > acquire the national boundary map for the Great lakes states in
> Africa.
>
> - John

--
Ansley Manke  Pacific Marine Environmental Laboratory  Seattle WA  (206)526-6246

Is anyone using Cave5d to visualize vis5d data?  If so, how are you
expanding the memory on Cave5d?  I have Cave5d version 1.4 (only the
executable) and I cannot change the default's memory.  Cave5d version
1.4 has 2 memory parameters that can be set to define the memory
(C5D_MEM_SIZE and VIS5D_MEM_SIZE), however cave5d ignores these
paramters when I set them in my config file.
Also, does anyone know where I can obtain the source code for cave5d?

Rashaad Jones
Research Assistant
Applied Research Laboratory
Pennsylvania State University

Hi Steven,

This is really great. Obviously I'm not doing much with
Vis5D recently, and I really appreciate the fine work
you've been doing with Vis5d+.

Cheers,
Bill

> I'm pleased to announce that revision 1.1 of Vis5d+ (release notes
> attached below) is available from the Vis5d+ web site:
> 
>                 http://vis5d.sourceforge.net/
> 
> Vis5d+ is intended as a central repository for enhancements to Vis5d.  It
> started out as a project to port Vis5d to the automated autoconf/automake
> build system, and has since benefited from a number of other enhancements.
> Vis5d+ also continues to track all changes in the mainline Vis5d sources.
> 
> The current release includes a number of bugfixes, enhancements to
> v5dimport, and now cleanly separates the API from the GUI.  (See below.)
> 
> Thanks especially to Jim Edwards for his continuing efforts on Vis5d+.
> He has also begun work on an alternative Gtk+/GNOME based graphical
> interface.  (Gtk+ is a popular GUI library for GNU/Linux and other Unices,
> and has even been ported to Windows.)
> 
> We welcome your comments.
> 
> Cordially,
> Steven G. Johnson
> 
> PS.  We also hope that someone will be motivated to merge the NCAR
> extensions into Vis5d+ (hint, hint)...
> 
> Vis5d+ 1.1
> 
>   * The API (i.e. the libvis5d library) is now properly independent of the GUI.
> 
>   * Better GrADS support in v5dimport (case-insensitive keywords, support
>     for the PDEF statement, and template and byteswapped options).
>     Removed an abitrary restriction on the minimum size of a file to be
>     imported, which caused certain small GrADS header files to fail.
> 
>   * v5dimport is now built using the same sources as the "import" command
>     in Vis5d, in the src/ directory (the import/ directory is removed).
> 
>   * Build maketopo and some other example utilities in the utils/ directory
>     by default, although they are still not installed since you need to
>     edit the source code for them to be useful.  Thanks to Markus
>     Neteler for the suggestion.
> 
>   * v5dimport now looks for EARTH.topo in the installation prefix, or
>     in the VIS5D_PATH/VIS5D_TOPO environment variables if one of those
>     are defined.
> 
>   * Synchronized with mainline tree as of April 2, 2001.
> 
>   * You can now optionally specify the year with 4 digits, and the
>     weekday is now correct.  Also fixed year 2100 problem (not a
>     leap year).  Thanks to Jaakko Hyvatti for the patch.
> 
>   * Added fortran-callable functions V5DUPDATE and V5DUPDATETIMES,
>     similar to the C interface; thanks to Jaakko Hyvatti.
> 
>   * Eliminate gcc compiler warnings; many code cleanups.
> 
>   * Fixed non-working --with-memory configure option, thanks to Andrew Woolf.
> 
>   * Fixed some problems compiling the Fortran utilities under Tru64 Unix,
>     reported by Frederic Chagnon.
> 
>   * Fixed possible crashing bug in vertical levels interpolation code.
> 
>   * Possible bug fix when reading profiles from NetCDF files.
> 
>   * Fixed a bug in v5dimport so that it properly handles the lowest level
>     setting in the vertical coordinate system.  For example, if a vcs
>     had 4 levels 1012 1000 750 500 and a variable had lowlev=1 Nl=3,
>     that variable should be defined on levels 1000 750 and 500, but
>     v5dimport was ignoring the lowlev and using levels 1012 1000 and
>     750.

-- 
----------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hi...@fa...  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html

Dear Vis5d users and developers,

I'm pleased to announce that revision 1.1 of Vis5d+ (release notes
attached below) is available from the Vis5d+ web site:

		http://vis5d.sourceforge.net/

Vis5d+ is intended as a central repository for enhancements to Vis5d.  It
started out as a project to port Vis5d to the automated autoconf/automake
build system, and has since benefited from a number of other enhancements.  
Vis5d+ also continues to track all changes in the mainline Vis5d sources.

The current release includes a number of bugfixes, enhancements to
v5dimport, and now cleanly separates the API from the GUI.  (See below.)

Thanks especially to Jim Edwards for his continuing efforts on Vis5d+.  
He has also begun work on an alternative Gtk+/GNOME based graphical
interface.  (Gtk+ is a popular GUI library for GNU/Linux and other Unices,
and has even been ported to Windows.)

We welcome your comments.

Cordially,
Steven G. Johnson

PS.  We also hope that someone will be motivated to merge the NCAR
extensions into Vis5d+ (hint, hint)...

Vis5d+ 1.1

  * The API (i.e. the libvis5d library) is now properly independent of the GUI.

  * Better GrADS support in v5dimport (case-insensitive keywords, support
    for the PDEF statement, and template and byteswapped options).
    Removed an abitrary restriction on the minimum size of a file to be
    imported, which caused certain small GrADS header files to fail.

  * v5dimport is now built using the same sources as the "import" command
    in Vis5d, in the src/ directory (the import/ directory is removed).

  * Build maketopo and some other example utilities in the utils/ directory
    by default, although they are still not installed since you need to
    edit the source code for them to be useful.  Thanks to Markus
    Neteler for the suggestion.

  * v5dimport now looks for EARTH.topo in the installation prefix, or
    in the VIS5D_PATH/VIS5D_TOPO environment variables if one of those
    are defined.

  * Synchronized with mainline tree as of April 2, 2001.

  * You can now optionally specify the year with 4 digits, and the
    weekday is now correct.  Also fixed year 2100 problem (not a
    leap year).  Thanks to Jaakko Hyvatti for the patch.

  * Added fortran-callable functions V5DUPDATE and V5DUPDATETIMES,
    similar to the C interface; thanks to Jaakko Hyvatti.

  * Eliminate gcc compiler warnings; many code cleanups.

  * Fixed non-working --with-memory configure option, thanks to Andrew Woolf.

  * Fixed some problems compiling the Fortran utilities under Tru64 Unix,
    reported by Frederic Chagnon.

  * Fixed possible crashing bug in vertical levels interpolation code.

  * Possible bug fix when reading profiles from NetCDF files.

  * Fixed a bug in v5dimport so that it properly handles the lowest level
    setting in the vertical coordinate system.  For example, if a vcs
    had 4 levels 1012 1000 750 500 and a variable had lowlev=1 Nl=3,
    that variable should be defined on levels 1000 750 and 500, but
    v5dimport was ignoring the lowlev and using levels 1012 1000 and
    750.

Hi

> Am currently running vis5d to display some weather model output over
Africa.
> The main problem that I have is that I can only overlay the continetal

> boundary
> and not the country's natioanl boundary map. Please assist me on how I
can
> acquire the national boundary map for the Great lakes states in
Africa.

- John