(NVIDIA's CUDA official home page: nvidia/object/cuda_home_cn.html
51CTO's Kaiyong Blog: openhero.blog.51cto/
CUDA? is a general-purpose parallel computing architecture introduced by NVIDIA that enables GPUs to solve complex computing problems. It includes the CUDA Command Set Architecture (ISA) and the parallel computing engine inside the GPU. Developers can now write programs for the CUDA? architecture using C, the most widely used high-level programming language. Programs written can then run with super high functionality on a CUDA?-capable manager. Other languages ??will be supported in the future, including FORTRAN and C .
The effect of CUDA technology is used to perform scientific/scientific operations. Although the office ability of CPU is very powerful, but because of the versatility of CPU, it integrates logic control, logic operation and other units, resulting in its low efficiency in numerical operation. The GPU itself is born for computing, because graphics management requires a lot of computing power. The design of the GPU is different from that of the general-purpose CPU. The GPU is a stream manager, so it is dozens or hundreds of times stronger than the CPU in terms of computing power. To exaggerate, a certain scientific/scientific computing problem may take a month on a general PC, but only three hours on a GPU (quoted from a lecture by a professor).
lecture6floating-point2008.ppt
lecture7casestudyvmd2008.ppt
lecture8casestudy2008.ppt
lecture9cudaconclusion2008.ppt)
收藏
