A Research Of Universal Quantum Computing Simulation In Heterogeneous Cluster Environment

In recent years,the great progress of quantum computing has attracted wide attention,quantum computers can solve some NP hard problem of classical computers in linear time due to its inherent parallelism,the potential of quantum computing has emerged although the number of existing quantum algorithms that outperforms the classical calculation are still small.However,there is still far way from the quantum computers be available.The physical implementation of quantum computer is still in small scale.Therefore,the quantum computer simulation technology is still an effective tool for validation of quantum algorithms.In this paper,the research on the application of CPU+GPU heterogeneous cluster for general quantum computer simulation is carried out.The main contents of this paper are as follows:1.We analyzed the communication characteristics of quantum computation simulation on heterogeneous cluster.To solve the problem of expensive communication cost,several optimization methods are proposed,which can reduce the communication cost in the process of quantum computation simulation effectively.2.In combination with the features of CUDA architecture,Kepler architecture and quantum computer simulation,the storage optimization and performance improvement of quantum computer simulation are carried out.3.Based on the parallel feature analysis of quantum computation,we build a CPU+GPU heterogeneous cluster architecture for general quantum computation simulation,several experiments are carried out to verify the effectiveness of the proposed method,and the performance of the simulation platform is analyzed.Based on the research above,with architecture design and performance optimization,we implemented a large-scale universal quantum computer simulation platform based on CPU+GPU heterogeneous cluster,and provide an effective tool for the research of quantum algorithm.