The GPU parallel algorithm of whole ordinal in universal combinatorics coding

Universal combinatorics coding is a new kind of coding method and whole ordinal plays an important role in universal combinatorics coding. Whole ordinal GPU parallel algorithm is proposed in this paper. Whole ordinal computing is implemented by combining the advantage of GPU in aspect of complicated and vast parallel computing with CPU technology, and employing CPU+GPU heterogeneous program development model. It makes whole ordinal parallel computing implement by calling kernel function of large number multiplication calculation in original serial program of computing whole ordinal. In the process of research, a series of GPU parallel problem are solved by optimizing program code many times. Experiments show that GPU parallel method improves the speed of calculating whole ordinal greatly, and then improves the computation efficiency of relevant experiments. Whole ordinal parallel implementation also has important meaning to estimating relevant experimental data and pushes the pace of universal combinatorics coding into practical at the same time.