Project 2 Proposal

Overview

For my second project, I would like to continue down the road of multiphase flow data visualization that I embarked upon in my first project. The previous project did not quite come to fruition as the final visualization as Paraview lacked some critical features. In particular, I would like to get the following Paraview features working, likely by applying the existing CCV patch and working with Ella Peterson to get a later version of Paraview (post 5.6.0-most likely).

• 1. Volume rendering

  2. Better FPS in the Yurt for a more immersive experience (there is likely an existing fix for this)

  3. Head-tracking

Once these things are implemented, I will hopefully be able to use datasets of the same type as my first project (multiphase flow data) in order to get a better experience visualizing the structures of interest in these data.

Calendar

This calendar is subject to change. I am not quite sure what exactly realistic and reasonable milestones will be with the transition to remote. 

• Some weekly goals:

  • 4/7: Build Paraview, following up the instructions with Ella and Camilo. Run some test data in the scalable build to ensure that this patch is running smoothly.

  • 4/9: Boot up the Yurt remotely and run the scalable build.  

  • 4/14: View multiphase flow data in the Yurt

  • 4/16: Coordinate with on-campus people to figure out how testing head tracking will work

  • 4/21: Figure out and try to install a version of Paraview that will allow for off-axis rendering. Have gone through the source, as well as taken steps towards figuring out how to get this distributed across the various machines in the Yurt. 

  • 5/5: Get the remaining deliverables going!

Accompanying Deliverables

• Instuctions on how to buld from scratch the newer Paraview version

• Comparison as to the utility of head tracking vs. not on the wiki

• Writeup with collabator on how improvements in Paraview were helpful

• Update Yurt Paraview megathread with new startup information

• Working user-ready build of a newer version of Paraview available in the Yurt

Class Activities

• Using Paraview in the Yurt remotely

• Trying the build process (ccmake or something, not sure if this is useful)

• Having people bring in their own datasets to Paraview and rendering surfaces for these

project 2 draft eval by David 3/31/20

Awesome set of 3 goals!  I'm worried that your plan/schedule/calendar is pretty sparse, but I understand that there is a lot of uncertainty.  I would suggest creating some more concrete deadlines, even if you are not confident of them.  They can be adjusted as more info comes to light.  As you make them concrete, you may also find some implicit dependencies that will add to your list of concrete things that need to happen.

There is more info on  volume rendering in the yurt.  It needs both off-axis rendering, which is only in the newer versions of paraview, and a hack that causes paraview to distribute the volume software to all server machines.  The scalable port of paraview has not been made to work with versions beyond 5.6, as I think you know, and even getting the 5.6 version to work has gotten stuck.  Ella and I discussed that yesterday, so you might want to check in on where that is currently failing.

Project 2 Evaluation by Ross 03/21/20

Great deliverables! I agree with David that fleshing out your calendar with concrete goals is necessary. I like the three goals -- volume rendering would be particularly cool; however, I'm not sure how hard it would be to test head tracking remotely. Also, it might be helpful if you specified what optimizations you planning to make to improve Paraview's performance in the YURT; would this involve changing the Paraview source code or are there some more surface level optimizations that we can make? Nevertheless, good work!