projects::CUDA

CUDA is a technology that makes it possible to leverage the computational power of NVIDIA GPUs for general purpose computing. In Gentoo, support for CUDA is provided by the following packages:

  • dev-util/nvidia-cuda-profiler
  • dev-util/nvidia-cuda-sdk
  • dev-util/nvidia-cuda-toolkit
  • dev-python/pycuda

I'm mainly interested in applications of CUDA in scientific computing, in particular in the numerical solution of stochastic differential equations (SDEs) and Lattice-Boltzmann (LBM) simulations of fluids. The speed-ups possible to attain in problems from these fields are of the order of 10x - 100x.

CUDA::SDE

For more information about using CUDA for stochastic dynamics simulations, see my paper Accelerating numerical solution of Stochastic Differential Equations with CUDA with the accompanying code SDE on CUDA.

CUDA::sailfish

Sailfish is my implementation of a LBM solver using CUDA. The project is at an early experimental stage, but is growing steadily. The code is hosted at Gitorious.

Check out the Sailfish Reference Manual to get an idea of what the project is about and how to use it to run your own fluid simulations.

Screenshots:

lid-driven cavity flow past a plate