[Paraview] Paraview on SGI Altix 4000

Moreland, Kenneth kmorel at sandia.gov
Tue Sep 23 11:16:11 EDT 2008


I mean the vendor-specific graphics driver provided by your system.  Often
we find render specific problems that are attributed to bugs in the system
graphics drivers, which are typically updated at a much faster rate than the
OpenGL or GLX standard changes.

When we see a graphics problem like this, it is hard to determine the actual
cause.  Thus, the first course of action is usually to update the vendor's
graphics driver to rule that out.  If the problem persists, then we look
elsewhere.

-Ken


On 9/22/08 9:14 PM, "Biao She" <shebiao at gmail.com> wrote:

> Hi Ken.
> Which OpenGL verison does paraview need? Also, by saying OpenGL
> drivers, do you mean the GLX version of the supercomputer? Coz my
> paraview server is running on a linux machine. The version of GLX on
> the machine is 1.2. The latest GLX version is 1.4. Do I need to update
> the glx to 1.4? Thanks!
>
> Aaron
>
> On Mon, Sep 22, 2008 at 11:51 AM, Moreland, Kenneth <kmorel at sandia.gov> wrote:
>> It sounds like you are having a framebuffer readback issue.  Can you check
>> to see if your OpenGL drivers are up to date?
>>
>> -Ken
>>
>>
>> On 9/22/08 11:39 AM, "Biao She" <shebiao at gmail.com> wrote:
>>
>>> Thanks for you reply, Ken.
>>> In stead of black square, the result is even more strange if I run
>>> without the --use-offscreen-rendering flag. It is basically a huge
>>> collection of random dots, and if I move the camera, the glxgears even
>>> appears in the display.
>>>
>>> Aaron
>>>
>>> On Mon, Sep 22, 2008 at 11:33 AM, Moreland, Kenneth <kmorel at sandia.gov>
>>> wrote:
>>>> It sounds like it could be a rendering problem.  Perhaps the offscreen
>>>> rendering is misbehaving.  What happens when you run without the
>>>> --use-offscreen-rendering flag?
>>>>
>>>> -Ken
>>>>
>>>>
>>>> On 9/22/08 11:28 AM, "Biao She" <shebiao at gmail.com> wrote:
>>>>
>>>>> Hi folks.
>>>>> Has anyone ever installed paraview on SGI Altix 4000? If I compiled
>>>>> paraview with "PARAVIEW_USE_MPI" option setting to ON, the paraview
>>>>> didn't work properly. For example, I first started the MPI version
>>>>> pvserver on the supercomputer by issuing "DISPLAY=:0 mpirun -np 4
>>>>> ./pvserver --use-offscreen-rendering",  then I launched paraview
>>>>> locally and connected it to the pvserver on the supercomputer.
>>>>> Normally, at the center of paraview display area, there should have
>>>>> three reference lines of a coordinate system. However, my paraview
>>>>> just displayed a black square.
>>>>> I have also tried the non-MPI version and the precompiled version of
>>>>> Paraview. They worked fine. I don't know if the problem is MPI
>>>>> implementation problem. I have heard that SGI's MPI implementation is
>>>>> kind of different. However, other programs, such as glxgears, worked
>>>>> fine by using the SGI implementation. Also, if I use the x forwarding
>>>>> to start the MPI version of pvserver, the rendering result is also
>>>>> correct. So, I am not quite convinced that the problem is MPI problem
>>>>> and I am quiet confused.
>>>>> I will highly appreciate any of your comments.
>>>>> Thanks!
>>>>>
>>>>> Aaron
>>>>> _______________________________________________
>>>>> ParaView mailing list
>>>>> ParaView at paraview.org
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>>
>>>>
>>>>   ****      Kenneth Moreland
>>>>    ***      Sandia National Laboratories
>>>> ***********
>>>> *** *** ***  email: kmorel at sandia.gov
>>>> **  ***  **  phone: (505) 844-8919
>>>>    ***      fax:   (505) 845-0833
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>>   ****      Kenneth Moreland
>>    ***      Sandia National Laboratories
>> ***********
>> *** *** ***  email: kmorel at sandia.gov
>> **  ***  **  phone: (505) 844-8919
>>    ***      fax:   (505) 845-0833
>>
>>
>>
>>
>


   ****      Kenneth Moreland
    ***      Sandia National Laboratories
***********
*** *** ***  email: kmorel at sandia.gov
**  ***  **  phone: (505) 844-8919
    ***      fax:   (505) 845-0833





More information about the ParaView mailing list