Solutions For Optimizing The Data Parallel Prefix Sum Algorithm Using The Compute Unified Device Architecture
Ion Lungu, Dana-Mihaela Petroșanu, Alexandru Pîrjan · RePEc: Research Papers in Economics · 2011
In this paper, we analyze solutions for optimizing the data parallel prefix sum function using the Compute Unified Device Architecture (CUDA) that provides a viable solution for accelerating a broad class of applications. The parallel prefix sum function is an essential building block for many data mining algorithms, and therefore its optimization facilitates the whole data mining process. Finally, we benchmark and evaluate the performance of the optimized parallel prefix sum building block in CUDA.