Fast Gauss Transform on CUDA
- 26 Nov: Serial c++ version, with k-mean clustering
- 27 Nov: Port the current c++ version to gpu
- 28 Nov: Tree structure implementation and port it to gpu
- 29 Nov: Adaptive pmax, radius etc.
- Load balancing and gpu performance improvement
- First working prototype by 6 December 2014
- Fast computation of sums of Gaussians in high dimensions, Raykar et al. [pdf]