NVIDIA GPUs execute groups of threads known as warps in SIMT (Single Instruction, Multiple Thread) fashion. Many CUDA programs achieve high performance by… 16 MIN READ

In this post we introduce the "register cache", an optimization technique that develops a virtual caching layer for threads in a single warp. It is a software… 16 MIN READ

In efficient parallel algorithms, threads cooperate and share data to perform collective computations. To share data, the threads must synchronize. 16 MIN READ

The CUDA 9 release includes support for Volta GPUs, Cooperative Groups programming model extensions, faster libraries, and improved developer tools. 17 MIN READ