Parallel Rendering Mechanism for Graphics Programming on Multicore Processors

摘要

The present day technological advancement has resulted in multiple core processors coming into desktops, handhelds, servers and workstations. This is because the present day applications and users demand huge computing power and interactivity. Both of these reasons have resulted in a total design shift in the way the processors are designed and developed. However the change in the hardware has not been accompaniedwith the change in the way the software has to be written for these multicore processors. In this paper, we intend to provide the integration of OpenGL programs on a platform which supports multicore processors. The paper would result in clear understanding how graphics pipelines can be implemented on multi-core processors to achieve higher computational speeds up with highest thread granularity. The impacts of using too much parallelism are also discussed. An OpenMP API for the thread scheduling of parallel task is discussed in this paper. The tool Intel VTune Performance Analyzer is used to find the hotspots and for software optimization. Comparing both the serial and parallel execution of graphics code shows encouraging results and it has been found that the increase in frame rate has resulted due to parallel programming techniques.