Background: This study implemented an algorithm geometrically interpolating a given polygon using uniform B-splines with parallel computing within GPU by OpenCL. Methodology: Parallel matrix calculation and multi-partition methods were proposed to accelerate computing ability. Results: The results showed the implementation within GPU had greater advantage in computing than CPU. Conclusion: The GPU?s implementation took less extra cost for improving the accuracy than CPU?s. The GPU had great value on high precision applications with numerous interpolated points.
展开▼
