Shelves

GPU architecture and parallel programming from the ground up: the CUDA execution model, memory hierarchy and locality, and the parallel patterns that recur everywhere (convolution, stencil, histogram, reduction, scan, merge, sort, sparse matrix computation, graph traversal), with case studies in deep learning and scientific computing.

Notes to come as I work through it.