Method and system of rendering parallel global illumination
A method of rendering parallel global illumination for obtaining a photo-realistic rendering image, includes: generating candidate parallel random numbers for creating a sample; creating a sample by selecting one of the generated random numbers; performing Monte Carlo integration on the basis of the created sample; and providing an application programming interface (API) common to operations having attributes of algorithm parallelism and data-level parallelism according to processed works, so that a common call interface calls a sample creating function, a random number generating function, and a Monte Carlo integration performing function. Furthermore, the method includes: demanding an operation regarding a processed work from a system call controller by the common call interface; performing parallel processing through a preset scheduling when the system call controller performs the operation demanded by the common call interface; and performing the processed work transferred from the system call controller by an operation hardware.
1. A method of rendering parallel global illumination for obtaining a realistic rendering image, the method comprising:
generating candidate parallel random numbers for creating a sample;
creating a sample by selecting one of the generated random numbers;
performing Monte Carlo integration on the basis of the created sample;
providing an application programming interface (API) common to operations having attributes of algorithm parallelism and data-level parallelism according to processed works, so that a common call interface calls a sample creating function, a random number generating function, and a Monte Carlo integration performing function;
demanding an operation regarding a processed work from a system call controller by the common call interface;
performing parallel processing through a preset scheduling according to the attribute of the processed work when the system call controller performs the operation demanded by the common call interface; and
performing the processed work transferred from the system call controller by an operation hardware.
2. The method of claim 1 , wherein the operation hardware comprises:
a multi-core central processing unit (CPU) in which a plurality of cores are mounted to one processor;
a graphic processing unit to which a programmable shader processor is mounted; and
a single instruction multiple data (SIMD) dedicated hardware processing a large amount of parallel data by using an array processor.
3. The method of claim 2 , wherein performing parallel processing through a preset scheduling further comprises:
instructing the multi-core CPU to process the work by the system call controller when the attribute of the processed work is a parallel algorithm work.
4. The method of claim 2 , wherein performing parallel processing through a preset scheduling further comprises:
instructing the dedicated SIMD hardware to process the work by the system call controller when the attribute of the processed work is a parallel data work and the amount of the processed data is larger than a preset amount of data.
5. The method of claim 2 , wherein performing parallel processing through a preset scheduling further comprises:
instructing the GPU to process the work by the system call controller when the attribute of the processed work is a parallel data work and the amount of the processed data is equal to or smaller than a preset amount of data.
6. A system of rendering parallel global illumination, the system comprising:
a parallel random number generator generating candidate parallel random numbers in order to create a sample for rendering of global illumination for obtaining a realistic rendering image;
a sample creator creating a sample by selecting one of the generated random numbers;
a Monte Carlo integration handler handling Monte Carlo integration on the basis of the created sample;
a common call interface calling functions of the sample creator, the parallel random number generator, and the Monte Carlo integration handler, providing an application programming interface (API) common to operations having attributes of algorithm parallelism and data-level parallelism according to processed works, and demanding an operation regarding a processed work from a system call controller;
a system call controller performing parallel processing through a preset scheduling according to the attribute of the processed work when the system call controller performs the operation demanded by the common call interface; and
an operation hardware performing the work transferred from the system call controller.
7. The system of claim 6 , wherein the operation hardware comprises:
a multi-core central processing unit (CPU) in which a plurality of cores are mounted to one processor;
a graphic processing unit(GPU) to which a programmable shader processor is mounted; and
a dedicated single instruction multiple data (SIMD) hardware processing a large amount of parallel data by using an array processor.
8. The system of claim 7 , wherein the system call controller instructs the multi-core CPU to process the work when the attribute of the processed work is a parallel algorithm work.
9. The system of claim 7 , wherein the system call controller instructs the dedicated SIMD hardware to process the work when the attribute of the processed work is a parallel data work and the amount of the processed data is larger than a preset one.
10. The system of claim 7 , wherein the system call controller instructs the GPU to process the work when the attribute of the processed work is a parallel data work and the amount of the processed data is equal to or smaller than a preset one.