Middleware for programming NVIDIA GPUs from Fortran 9X
Nail A. Gumerov, Ramani Duraiswami, W. Dorland · 2007
Over the past year, the major graphics vendors have realized there is a market for these in HPC, and have started producing tools to allow programming them from high level languages, ATI’s CloseTo-Machine (CTM) and NVIDIA’s CUDA [1]. We are more familiar with the latter architecture, for which both a beta programming environment and publicly available graphics boards with impressive speedups over their previous generation, and restrict ourselves to them.