Cuda time is not long, the first is in the Cuda-convnet code to contact Cuda code, it did look more painful. Recently Hollow, the library has borrowed this GPU high-performance programming Cuda combat to see.
Jeremy Lin
What is Cuda
CUDA (Compute Unified Device Architecture) is a computing architecture designed specifically to improve the efficiency of parallel program development. The CUDA architecture maximizes the computing power of the GPU when building high-performance applications. For more information, refer to Nvidia's about PAGE.
CUDA Development Environment Configuration
Before we get started, we've configured the CUDA development environment.
(1) The first step is to determine the computer with a CUDA-enabled GPU.
You can view it by right clicking on My Computer - management - Device Management - display adapter
My computer on this GPU is rotten, is GeForce GT 620, but fortunately, it supports CUDA. Other Cuda-enabled GPUs you can view via Nvidia's website:Https://developer.nvidia.com/cuda-gpus
(2) Install the GPU device driver (normally, this is already installed)
(3) Download CUDA Development Toolkit
You can download from https://developer.nvidia.com/cuda-downloads to the appropriate Cuda Toolkit.
I downloaded the cuda_6.5.14_windows_general_64.exe. The installation process is a fool process, always next and OK.
(4) Verify the installation is correct
This is also very simple, open cmd.exe on the computer, and then type nvcc-v
Then, change the path into Cuda samples to run DeviceQuery.exe. The possible paths in Windows 7, Windows 8.1 are:
C:\ProgramData\NVIDIA Corporation\cuda Samples\v6.5\bin\win32\release
The path on my computer is:
The results of the operation are as follows:
(5) New Cuda program
Open: Visual Studio-FILE-New-project-Nvidia-cuda 6.5
This address: http://blog.csdn.net/linj_m/article/details/41314763
More resources please follow blog: linjm-machine vision Weibo: Lin Jianmin-Machine Vision
Cuda Learning log: Windows Development environment Configuration