More and Faster: glInterleavedArrays glDrawArrays

Processing Coding Questions
1

Anyone have pointers as to how I might move toward being able to use the array functions from openGL (glInterleavedArrays and glDrawArrays) within processing? Am ok if this involves my needing to develop a library. Google searches on processing “glInterleavedArrays” returned nothing.

Initial application space is converting my present capability to show CNC tool paths. This is a full 3D slider<==>time pseudo VCR of the developing trenches in the material as the bit passes through. Current code is in C++. Used the glInterleavedArrays and glDrawArrays techniques to avoid calls to methods for every point in the fairly extensive TRIANGLE_STRIP’s.

My Background: Started as a course III. Wrote a 3D fly-thru openGL C++ app for all 100K wires in an IBM mainframe. Wanted to do web stuff so wrote http://www.webgltutorials.appspot.com/WebGLTutorial06 in java and javascript with full homebrew GL_Shaders. Think processing hits the sweet spot on GUI, adaptability, community, platform access, and extendability.

Hope to help.

1 Like
2

@TPMoyer – this sounds great. You should also post a hello in Please Introduce Yourself

IANA expert in openGL, but we do have a few on the forum – for example, I believe @gohai is quite experienced with this and has written Processing + openGL libraries. Hopefully someone can suggest starting points.

3

I wish I knew anything about GL - I am just an occational plumber.

I would suggest looking directly at the source code of Processing’s P2D/P3D renderer, to see if it provides access to what you want to use. You find the files here. Compiling and runnig a modified version is also surprisingly easy, as you only need a recent JDK and ant installed (the a link in the Readme on Processing’s Github project has more information about that).

The linked P2D/P3D renderer makes use of the JOGL library, which makes it possible to call the GL C API from within Java. So you could attempt to add missing functionality by adding the necessary methods or classes in P2D/P3D.

1 Like
4

beginPGL probably gives access to the required functionality, but if not you can directly access JOGL from it - see https://github.com/processing/processing/wiki/Advanced-OpenGL

It’s quite easy to extend the P2D/P3D renderers to add additional functionality without having to fork core Processing.

It’s also worth looking at PShape with the GL renderers, or at least the code of.

5

Good pointers!
Got this up and working by riffing off the examples/Demos/Graphics/LowLevelGLVboInterleaved.
Shared this as github https://github.com/TPMoyer/Processing3-using-low-level-OpenGL-VBO.

Did not need to recompile the Processing source.

The prior example showed how to use the beginPGL structures to
invoke shaders,
set up both the to-be-drawn data and the shader variables
push data to the GPU
launch a single method DrawArrays() call to roll through all the data.

How might one nominate a sketch for inclusion in the distributed examples?

The present code should enable substantially faster rendering than the prior published example. It avoids the needless re-calculation and of the dataset for each Draw(), and avoids the needless data movement from the CPU to the GPU for each Draw().

1 Like
6

Regarding distributed examples, I believe you can just put the sketch in a pull request to the examples repository.

I believe this (with the xml index) is used to build both the examples in PDE and the examples on the website.