Fast parallel interpolation algorithm using CUDA

Interpolation is one time consuming and serial operation in the fields of spatial information processing. While fast processing speed is often required in some real-time interactive scenarios. With the development of Purpose computing on Graphics Processing Units (GPGPU), it provides an opportunity to accelerate some traditional inefficient algorithms with low-cost and low-power compared to clusters. In this paper, we mapped the Inverse distance weighted (IDW) interpolation method to Compute Unified Device Architecture (CUDA) parallel programming model. Taking the advantage of Graphics Processing Unit (GPU) parallel computing, we build two-level indexes on GPU, then clever blocking schemes are used to assign computing task among different threads. After illustrate the parallel interpolation process, we conduct several experiments, the result shows the correctness and high efficiency of our optimized implementation. With larger influence radius and massive data, the performance can obtain dozens of times speedups over a very similar single-threaded CPU implementation.