May 2019
Intermediate to advanced
452 pages
12h 16m
English
Every NVIDIA GPU has its own general specifications and available features. These parameters can be understood better through compute capability, which is a standard set by NVIDIA to classify each GPU according to its hardware and software specifications/features. Through compute capability, NVIDIA GPUs can be compared easily based on such classifications. As the CUDA API evolves, so does the compute capability. The most recent CUDA SDK 10.1 support for compute capability is from 3.0 to 7.5 for NVIDIA GPUs belonging to Kepler, Maxwell, Pascal, Volta, and Turing architectures. More information on this can be found at the following link: https://developer.nvidia.com/cuda-gpus.
Read now
Unlock full access