.PNG)
CUDA
Sign in to saveAlso known as Compute Unified Device Architecture, Nvidia CUDA, Nvidia Compute Unified Device Architecture, NVIDIA CUDA, NVIDIA Compute Unified Device Architecture, Nvidia Cuda, NVIDIA Cuda
NVIDIA所推出的一種軟硬體整合技術
In the Vinony graph
Within Vinony's link graph, CUDA is referenced by 469 other articles, and connects out to shader, CPU cache and general-purpose computing on graphics processing units.
Vinony files it under Computer physics engines, GPGPU and GPGPU libraries.
Its subject is documented across 36 Wikipedia language editions.
Key facts
- Software.name
- CUDA
- Software.screenshot
- Nvidia CUDA Logo.jpg
- Software.author
- Ian BuckJohn Nickolls
- Software.developer
- Nvidia
- Software.programming language
- C
- Software.operating_system
- Windows, Linux
- Software.platform
- Supported GPUs
- Software.genre
- GPGPU
- Software.license
- Proprietary
via Wikipedia infobox
Wikidata facts
- Instance of
- application programming interface
- Developer
- Nvidia
- Official website
- developer.nvidia.com/cuda-zone
- Image
- Nvidia CUDA Logo.jpg
- Has use
- concurrent computing
Show 6 more facts
- software version identifier
- 13.2.0
- inception
- 2007-06-23
- Commons category
- CUDA
- operating system
- macOS
- copyright license
- proprietary license
Sources (6)
via Wikidata · CC0
Article · 中文
CUDA（Compute Unified Device Architecture，统一计算架构）是由英伟达NVIDIA所推出的一種整合技術，是該公司對於GPGPU的正式名稱。透過這個技術，使用者可利用NVIDIA的GPU进行图像处理之外的運算，亦是首次可以利用GPU作為C-编译器的开发环境。CUDA 開發套件（CUDA Toolkit ）只能將自家的CUDA C-語言（對OpenCL只有链接的功能），也就是執行於GPU的部分編譯成中間語言或是特定NVIDIA GPU架構的機器碼（NVIDIA 官方稱為 "device code"）；而執行於中央处理器部分的C / C++程式碼（NVIDIA 官方稱為 "host code"）仍依賴於外部的編譯器，如Microsoft Windows下需要Microsoft Visual Studio；Linux下則主要依賴於GCC。
Abstract from DBpedia / Wikipedia · CC BY-SA