Yes, it is possible though not necessarily straightforward, depending on how much visualization algorithm and Cuda/GLSL experience you have. Also, you will have to write a new mapper since the current one works on data on main memory. If I was estimating how much this would take, I would estimate no less than 6 months for something like 4 algorithms and 1 mapper.
On Fri, Feb 26, 2010 at 3:35 AM, liuning <[email protected]> wrote: > Hi Berk, > > Thank you for your reply. One of the priorities of our in-situ > visualization work is speed, so moving data to main memory may not be a > preferred approach. We just need several basic visualization functions(e.g. > slice and contour ) and analysis functions(e.g. sub-sampling ). Is it > possible that we write our own filters running on GPU and integrate them > into the current VTK pipeline? Since the sources, readers,writers and > filters are configured via XML file, if we set up the appropriate filter > XML configuration file, will it work for us ? > > Thanks again, > > -Ning > > > On Thu, Feb 25, 2010 at 10:29 PM, Berk Geveci <[email protected]>wrote: > >> It depends on how you want to do it. If you want to keep the data on the >> GPU memory and perform all of the analysis/visualization on the GPU, VTK is >> currently not the right tool. Almost all of VTK's filters work on the CPU >> whereas rendering happens on the GPU. If you are OK with moving data to main >> memory whenever you want to do analysis/vis, the in-situ vis work we are >> doing would be a fit for you. Let me know if you are interested and I can >> get you started. >> >> -berk >> >> On Wed, Feb 24, 2010 at 10:06 PM, liuning <[email protected]> wrote: >> >>> Hi all, >>> >>> We attempt to do a in-situ visualization to visualize the simulation, >>> which is mainly performed on GPUs. I noticed that there was a presentation >>> in Vis09 describing how to do in-situ visualization with the help of Python >>> scripts. Since the simulation data reside in GPUs, I wonder whether the >>> approach mentioned above can do this. Or, are there some better methods to >>> do that? >>> >>> Thanks a lot. >>> >>> -Ning >>> >>> >>> >>> _______________________________________________ >>> 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://www.paraview.org/mailman/listinfo/paraview >>> >>> >> >
_______________________________________________ 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://www.paraview.org/mailman/listinfo/paraview
