[Paraview] Client server volume rendering problem in 4.1 and 4.2RC1
ronald.fowler at stfc.ac.uk
ronald.fowler at stfc.ac.uk
Fri Sep 19 09:34:19 EDT 2014
Hi Utkarsh,
I have just downloaded the latest nightly build and it seems to work fine now (on Windows 7/64bit), doing remote volume rendering irrespective of the settings for remote/local. Many thanks for that quick fix.
Ron
-----Original Message-----
From: Utkarsh Ayachit [mailto:utkarsh.ayachit at kitware.com]
Sent: 17 September 2014 17:30
To: Fowler, Ronald (STFC,RAL,SC)
Cc: ParaView
Subject: Re: [Paraview] Client server volume rendering problem in 4.1 and 4.2RC1
Ron,
Here's a fix that makes ParaView use remote-rendering irrespective of remote-rendering threshold when volume rendering:
http://review.source.kitware.com/#/t/4672
It'd be great if you could give it a try to confirm that it works.
Thanks
Utkarsh
On Wed, Sep 17, 2014 at 10:09 AM, Utkarsh Ayachit <utkarsh.ayachit at kitware.com> wrote:
> Ron,
>
> Those are good observations. I has been on my todo list to clean up
> the remote-local rendering issue. I want ParaView to automatically
> switch to remote rendering, if possible when volume rendering. I'll
> see what I can do for that. The volume rendering code is currently
> being revamped. I'd expect better "smart" rendering in near future.
>
> Utkarsh
>
> On Wed, Sep 17, 2014 at 6:29 AM, <ronald.fowler at stfc.ac.uk> wrote:
>> Hi,
>>
>> A post by Robert Atwood in June mentioned a problem with client
>> server volume visualisation in PV4.1, and I have also seen problems
>> in this case both on Linux and Windows systems.
>>
>>
>>
>> Fixes (0014794 and 0014792) have been added to 4.2RC1 which allow
>> correct rendering of the wavelet test case. I tried this on a Windows
>> 7 system and works ok for 4.2RC1, but not for 4.1.
>>
>>
>>
>> However I think that the remote/local render problem for volume
>> visualisation is really very confusing and needs to be improved.
>> After spending more time then I should with it, I conclude that
>> (again on windows
>> 7 64 bit):
>>
>> (1) Volume rendering only works if done on the remote server, not the
>> local on (as acknowledged in a tracker ticket, though I have not seen
>> it documented elsewhere).
>>
>> (2) For 4.1 to do this correctly for just a volume image this seems to
>> need settings of “remote render threshold” ticked and “0 Mbytes” selected.
>> It will work for cases when >0 Mbytes is selected, if there is also
>> another object displayed (e.g. an isosurface). This led me to think
>> an additional surface was always needed to see the volume, which is not the case.
>>
>> (3) For 4.2RC1 the interface to settings has changed and there is no
>> longer a tick box. Instead the slider for remote local render size
>> has a much greater range. This should be set to zero to force remote
>> render for volume views. It also works for non-zero values less than
>> the size of the volume data, e.g. 4Mbytes for a 256^3 volume. This
>> setting did not seem work for 4.1 to force a remote rendering of just volume data.
>>
>>
>>
>> I think it would be useful if at least the tool tips in the settings
>> panel suggested how to make client server volume rendering work. It
>> would also be useful if a warning message were displayed when trying
>> (and failing) to do a local volume render in client server mode.
>>
>>
>>
>> Another problem that has been observed is that the “smart” option for
>> volume rendering is not necessarily that smart when the server nodes
>> have no GPUs and the client does. This was on a Linux 64 bit system with 4.1.
>>
>>
>>
>> Thanks,
>>
>> Ron fowler
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> --
>> Scanned by iCritical.
>>
>>
>>
>> _______________________________________________
>> Powered by www.kitware.com
>>
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>>
>> Please keep messages on-topic and check the ParaView Wiki at:
>> http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://public.kitware.com/mailman/listinfo/paraview
>>
More information about the ParaView
mailing list