基于cuda的itk-CUDAITK.pptVIP

CUDA ITK Won-Ki Jeong SCI Institute University of Utah NVIDIA G80 New architecture for computing on the GPU GPU as massively parallel multithreaded machine One step further from streaming model New hardware features Unified shaders (ALUs) Flexible memory access (scatter) Fast user-controllable on-chip memory Integer, bitwise operations NVIDIA CUDA C-extension NVIDIA GPU programming language No graphics API overhead Easy to learn Support development tools Extensions / API Function type : __global__, __device__, __host__ Variable type : __shared__, __constant__ cudaMalloc(), cudaFree(), cudaMem