VertexAPI2 is a CUDA/C++ template library for rapid implementation of algorithms for large-graphs, developed by Royal-Caliber for the DARPA XData program.
NVIDIA GTC 2014 Talk on VertexAPI2, an API for rapid development of large-graph analytics on the GPU, to be presented on Monday Mar 24, 11:00 AM in Room 210E.
NVIDIA GTC 2013 Talk on the use of advanced techniques for a highly optimized Smith-Waterman kernel for Kepler-class GPUs.
‣Moderngpu provides a very interesting survey of GPU techniques, a must-read for anyone working with GPUs/CUDA.
‣NVIDIA Cub is an excellent reusable template library of basic operations at the thread, warp and device level for CUDA development.
‣The Silicon Valley HPC & GPU group often has interesting talks and presentations by professionals across the GPU and HPC space. A good forum for any level of experience.