VMD-L Mailing List

From: Pawel Wielgus (wielgus_at_mml.ch.pwr.wroc.pl)
Date: Tue Mar 13 2007 - 08:07:38 CDT

Ha! It works! But just after I exported this long variable inside VMD
starting script. Anyway, so far looks stable, no problems.

Neverthless I still would have a few questions...how can one draw a
color scale-bar on the screen? And then adjust it (min and max values,
choosing colors and so on...
Any references?

Thanks for help!

Pawel

John Stone wrote:
> Pawel,
> Did you verify that the system variable is being detected by VMD
> at startup? What do your VMD startup messages say now? It could also
> simply be a bug in 3-D (volumetric) texturing in the ATI driver. Does
> color by "Volume" work for you, or does that similarly produced a solid
> black object? With VMD 1.8.5, the VolumeSlice representation should be
> relatively insensitive to various driver bugs because texture coordinates
> are explicitly calculated for the vertices of the VolumeSlice plane, so
> as long as the video driver does its job, I would expect you to see a
> correctly textured slide plane. If you change the background color
> to white, do you see a black slice plane, or do you see nothing at all?
>
> John Stone
> vmd_at_ks.uiuc.edu
>
> On Mon, Mar 12, 2007 at 09:00:45AM +0100, Pawel Wielgus wrote:
>
>>Dear John,
>>
>>after exporting the system variable, I can see sth like this:
>>
>>Info) File loading in progress, please wait.
>>Info) Using plugin cube for structure file test.cube
>>cubeplugin) trying to read cube data set 0
>>Info) Analyzing Volume...
>>Info) Grid size: 80x80x80 (7 MB)
>>Info) Total voxels: 512000
>>Info) Min: 0.000000 Max: 0.706192 Range: 0.706192
>>Info) Computing volume gradient map for smooth shading
>>Info) Added volume data, name=test.cube : Gaussian Cube: kom density=scf
>>Info) Using plugin cube for coordinates from file test.cube
>>Info) Determining bond structure from distance search ...
>>Info) Finished with coordinate file test.cube.
>>
>>Looks not bad. But that's all. I still can't see anything on my screen
>>when the VolumeSlice representation is chosen :(
>>
>>
>>Pawel
>>
>>
>>
>>John Stone wrote:
>>
>>>Pawel,
>>> Ah, I see the issue. Well, you have a choice of either force-enabling
>>>full rendering functionality (and risking crashes due to a number of
>>>known ATI driver bugs that are to my knowledge still not fixed based
>>>on recent reports from others), or using a different machine or a different
>>>video card. If you wish to take your chances with the ATI video driver
>>>and tell VMD to enable all rendering features, you can do this by
>>>setting the environment variable VMDDISABLEATILINUXWORKAROUND, like this:
>>> setenv VMDDISABLEATILINUXWORKAROUND 1
>>>
>>>After setting that environment variable, you should be able to start
>>>VMD again, but without the functionality limitations imposed by default
>>>when an ATI Linux driver is detected. Do not be surprised if you manage
>>>to crash VMD, the X server, or even lockup your machine, as I've seen all
>>>of these situations occur with both the open source and ATI provided
>>>drivers for Linux. Oddly, the same hardware works fine with the ATI
>>>provided Windows driver, so I think this is mainly an issue with their
>>>Linux driver series.
>>>
>>>If you're able to use VMD reliably with full features with your ATI
>>>Linux driver I would be very interested to hear about it, as many other
>>>ATI users have encountered stability problems when using the advanced OpenGL
>>>rendering features. Any tips that help others with their ATI drivers
>>>are appreciated.
>>>
>>>If you have stability problems when full functionality is enabled I would
>>>report this to ATI on their support forums. If you don't get a response
>>>from them, then it may be time to consider getting a different video card
>>>if working with volumetric data is important to you.
>>>
>>> John Stone
>>> vmd_at_ks.uiuc.edu
>>>
>>>On Fri, Mar 09, 2007 at 04:54:26PM +0100, Pawel Wielgus wrote:
>>>
>>>
>>>>Thanks for your reply.
>>>>When my VMD starts, one can read:
>>>>
>>>>Info) ATI Linux driver detected, limiting features to avoid driver bugs.
>>>>Info) OpenGL renderer: MOBILITY RADEON X300 Generic
>>>>Info) Features: STENCIL MDE CVA MTX PP PS GLSL(OVF)
>>>>Info) GLSL rendering mode is NOT available.
>>>>Info) Textures: 2-D (2048x2048), 3-D (2048x2048x2048), Multitexture (8)
>>>>
>>>>I really don't know what 'limiting features' mean. What I know for sure
>>>>is that I really had lots of troubles when compiling ATI fglrx drivers.
>>>>And I did it finally only against 2.6.10 kernel. But that's a long story...
>>>>
>>>>What do you think? Any special features in fglrx configuration needed ?
>>>>
>>>>
>>>>Pawel
>>>>
>>>>
>>>>
>>>>
>>>>John Stone wrote:
>>>>
>>>>
>>>>>Pawel,
>>>>> Yes, the VolumeSlice representation has been available for some time
>>>>>and should work fine so long as you have a video card that supports
>>>>>3-D (volumetric) texturing. You can check this by reading the VMD OpenGL
>>>>>related startup messages. You should see a texturing info line like
>>>>>the one I've pasted below:
>>>>> Info) OpenGL renderer: ZFB, VIS
>>>>> Info) Features: STENCIL MSAA(16) MDE MTX GAS
>>>>> Info) Textures: 2-D (4096x4096), 3-D (512x512x512), Multitexture (4)
>>>>>
>>>>>The important part of that status line is: "3-D (512x512x512)"
>>>>>
>>>>>What video chipset are you using? Have you checked to make sure
>>>>>that you're using the latest video driver available for your hardware?
>>>>>
>>>>> John Stone
>>>>> vmd_at_ks.uiuc.edu
>>>>>
>>>>>On Fri, Mar 09, 2007 at 04:27:17PM +0100, Pawel Wielgus wrote:
>>>>>
>>>>>
>>>>>
>>>>>>Dear VMD Users,
>>>>>>
>>>>>>may I ask you a question concerning the visualization of volumetric data?
>>>>>>I have a simple example of density for CH4 (just for graphical testing)
>>>>>>in a Gaussian03 cube format. The volumetric data is read properly, and
>>>>>>using 'Drawing method = Isosurface' I can easily get the density.
>>>>>>
>>>>>>But as only I switch to 'Drawing method = VolumeSlice' everything
>>>>>>disapears. As far as I know the volume slices (along cartesian axis) are
>>>>>>possible with VMD-1.8.5.
>>>>>>
>>>>>>What wrong am I doing ? Should I switch any options on ? Could you give
>>>>>>me any hint?
>>>>>>
>>>>>>
>>>>>>Regards
>>>>>>
>>>>>>
>>>>>>Pawel Wielgus
>>>>>>
>>>>>>
>>>>>
>>>>>
>