The Application Research Of Parallel Volume Rendering Based On Multi-Core CPU

Volume rendering is an important technique of visualization in scientific computing, and it can describe the interior structure information of 3D data object. But volume rendering is with large computing and long rendering time, and recently can't meet the requirement for real-time interactive analysis. So how to accelerate the process of rendering became an important goal of visualization reasearch.Rencently many accelerated methods for volume rendering have been suggested, included accelerating the process at the cost of image quality, depending on specical hardware and parallel computing on cosmical parallel machine. And parallel processing became the main research direction because of its flexibility and performances of high accellerating rate.Parallel processing includes parallel computer architechture, parallel algorithm and parallel programming language. Parallel processing combines parallel computer architechture and parallel algorithm together by parallel programming language. That is to make multi-proceesor perform one or more tasks synchronously based on right partition and assignment. Then the computing ability of the whole system is enhanced greatly. Also task partition of parallel volume rendering has two classes—object based and image based.This paper has two reserch parts. One discussed the parallel mechanism of VTK and its application on multi-core CPU. The other realized the OpenMP-based parallel ray casting algorithm based image partition on double-core computers. Experiments show that this method can improve the computing efficiency and is with pressing engineering application value and extending applicability.VTK is not only with powerful visualization function but also support parallel processing. Its compositive MPI and Sockets can afford facilities to the development of parallel visualization programming. Multi-core can make servers parallel and bring users powerful computing performance. Comparing with the single core, multi-core CPU brings us more powerful parallel processing ability, higher computing speed, lower time frequency and low cost. Meanwhile OpenMP standard which is an application programming interface (API) on parallel programming model of shared -memory is especially right to parallel programming on multi-core CPU.