Warning This document has been replaced by a new CUDA Programming Guide. The information in this document should be considered legacy, and this document is no longer being updated as of CUDA 13.0. Please refer to the CUDA Programming Guide for up-to-date information on CUDA. 1. Overview CUDA is a parallel computing platform and programming model developed by NVIDIA that enables dramatic increases

