SectionPlane cut

Greetings! Lately i used to do section Plane cuts, but due my vast geometry it obtains time consumation. Is there anyway to get it faster?

What graphics driver are you using? The OpenGL driver uses OpenGL clip planes, which should be fast, independent of the complexity of the geometry.

I have Ubuntu and installed all the modern OpenGL following this command:
sudo apt-get install libglu1-mesa-dev freeglut3-dev mesa-common-dev

also i have installed QT5

In Geant4 i was using OGL, OGLI, OGLIQt. The third one is slower in longitudinal and transverse cuts, but the first and second do longitudinal cut immidiately and transverse still slow.

Hi Alex

Mmm. The “I” in the driver name means “immediate”, which means the drawing goes immediately to the screen without creating a graphical database, which means every change of view parameters requires a “kernel visit”, i.e., a conversion from G4 objects (G4Box, etc.) to screen coordinates, which can be time consuming.

Equivalent drivers with a “S” (which means “stores”, which means G4 objects are converted to graphics primitives and stored in a graphical database) is usually much faster. It may take time to build the graphical database but it can be rendered from different points of view, different clip planes, etc., quickly.

The odd thing is that OGL usually translates to OGLSQt, so it should be quick. Was there a reason you chose OGLIQt? Let me see the list of registered drivers (usually in the first few lines of output).

Not sure what you mean, “longitudinal and transverse cuts”.


I had an idea if i use qt for programming as IDE then it will render my scene for sure.

OpenGLImmediateQt (OGLIQt, OGLI)

Rendering works on 1 core. Can i somehow give him more cores to work with? OR its not the case?
its breaks in speed for only one part, but when it hidden, rendering goes immidiately.

I meant, why did you choose OGLIQt rather than OGLSQt? Why “I” rather than “S”? “S” should work better. Just use OGL - we have programmed so that OGL gives you the best option in general.

Building the graphical database uses only one core (no easy way to parallelise it). Rendering the graphical database to the screen should exploit the graphics engine and thus should be fast (e.g., a change of viewpoint), so depends on what graphics card your computer has.

Not sure what you mean, “its breaks in speed for only one part, but when it hidden, rendering goes immidiately”?

I have an complicated part with a lot bool cuts , whereas other parts is relatively simple. Hidden is means i use /vis/geometry/set/visibility Name 0 false. So when my complicated part didn’t show on display process of rendering becomes much faster.