The Living Thing / Notebooks : GPU computation from the fringes.

Lesser-known options for doing SIMD computation with fewer tears for people, like me, who give not a damn about implementation details they, because just want it to work fast enough.

Aside: you’re going to have to mess around with downloading proprietary GPU toolkits from the manufacturer. Tedious. Don’t bother. Pay some cloud provider to rent theirs.

  1. Just writing GSL shaders using your compiler and the relevant manufacturer toolboxes. Laborious and tangential, unless you are a GPU-algorithm researcher. But could be fun I s’pose. See the book of shaders.
  2. for data-oriented computational data flow graphs, use one of those toolkits from the deep_learning community. These are easy and performant, although not quite as general as just writing a damn shader.

Neither fits?

OK, try these (emphasis on integration with python):

The book of shaders

This book will focus on the use of GLSL pixel shaders. First we’ll define what shaders are; then we’ll learn how to make procedural shapes, patterns, textures and animations with them. You’ll learn the foundations of shading language and apply it to more useful scenarios such as: image processing (image operations, matrix convolutions, blurs, color filters, lookup tables and other effects) and simulations (Conway’s game of life, Gray-Scott’s reaction-diffusion, water ripples, watercolor effects, Voronoi cells, etc.). Towards the end of the book we’ll see a set of advanced techniques based on Ray Marching.