How to increase the speed of volume rendering


I am new to Paraview and need some assistance in what I can do to reduce the time to do volume rendering.
Firstly my computer specs are:
AMD Ryzen 7 1800X
16 GB ram (going to be upgrading to 64 GB soon)
SSD hard drive
Radeon RX 480 8Gb graphic card.

I also have an Apple Mac bookpro (early 2013) with and i7, 16 GB ram, nvidia 650M running Catalina

So what I have done is run the motorBike openfoam case in version 6 and have attempted to do the volume rendering of the velocity of the air for the motorBikeLES results however it takes about a minute to do the rendering on my desktop computer running windows 10 in paraview version 5.7. haven’t tried on in ubuntu. The graphics card is also not showing about 5 % activity during the render while the CPU is running at max ( only one core as I was not running in MPI mode at the time but I will give it a try)

On my mac it can’t do the render in 5.7 however it can do it in 5.6 and it is faster than my desktop.

I have also tried the F-17 tutorial results on my mac and it works better in 5.6 than 5.7.

The motor bike data is 77.2 mb and the F17 data is 17.5 mb.
So I have a couple of questions that I need some assistance on.

  1. Is there a video/wiki/tutorial on how to improve the render time while working before a full render is required?
  2. Does the file format make a difference in the rendering time? So should I change the openfoam data to vtk or any other data type?
  3. In the F17 case when I use the extract block filter and extract the internal mesh and then change the representation to volume I can’t choose the volume rendering mode anymore as I can with the wavelet example. Then there is a an extracted block index box and a select mapper drop down menu with a a choice of “Project tetra, Z sweep, Bunky ray cast, Resample to Image”. On 5.7 only resample to image works all the others cause paraview to hang and I have to force quit. I am doing something wrong or is this a known issue with paraview?
  4. Should I buy a nvidia RTX card as it seems Kitware supports nvidia more than AMD?


I think I found my main problem that I was using an unstructured mesh/data. This video helped a lot It seems I need to do research to see it snappyhexmesh can create a structured mesh.
Resampling to image filter and then volume rendering is a lot faster.