![[PukiWiki]](image/bc_logo.png "[PukiWiki]"){kind=logo}
CUDA is a parallel computing platform and programming mode, which are developed by NVIDIA. This is implemented by graphics processing units. CUDA allows developers to access the virtual instruction set and memory of parallel computational elements in CUDA GPUs. Since the power of GPUs is allocated to general purpose processing, this provides more computational power to support its CPU.
CUDA Toolkit is an integrated development environment, which includes GPU compiler, libraries, drivers, and tools. This integrated development environment manages debug, optimization, and execution of an application.
CUDA is mainly based upon C/C++. For farther details of CUDA, please visit the URL shown below.
https://developer.nvidia.com/about-cuda
Most of GPU boards, which are NVIDIA produces such asTesla, Quadro, and GeForce, support CUDA. Recently, NVIDIA has introduced Jetson TK1, an ARM SOC board, which is equipped with Tegra K1. For more up-dated information, please visit the URL below.
https://developer.nvidia.com/cuda-gpus
This article introduces Jetson TK1, which is equipped with ARM SoC Tegra K1, and explains how to install L4T (Linux for Tegra) as well as CUDA 6 Toolkit from the scratch. This also lists the results of executing the sample programs. There is comparison between results of Jetson TK1 and Tesla K20c.
For more information, please visit this page.
This article explains how to install CentOS 6.4 and CUDA 5 on a PC which comes with Tesla K20c. This article shows the step-by-step instructions of how to set up CentOS as well as CUDA 5. It lists the results of the sample programs for CUDAs,
For more details, please go to this page.
- 2014/12/22 This article is initially published
