2018/01/01 by Ryan LaRose, LaRose, Ryan · 1 citation
Computer Science · #Distributed #Distributed and Parallel Computing Systems #FOS: Computer and information sciences #FOS: Physical sciences #Parallel #Parallel Computing and Optimization Techniques #Quantum Computing Algorithms and Architecture #Quantum Information and Cryptography #Quantum Physics (quant-ph) #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.1801.01037
openalex publication_date 2018/01/01 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
In this paper we describe, implement, and test the performance of distributed\nmemory simulations of quantum circuits on the MSU Laconia Top500 supercomputer.\nUsing OpenMP and MPI hybrid parallelization, we first use a distributed\nmatrix-vector multiplication with one-dimensional partitioning and discuss the\nshortcomings of this method due to the exponential memory requirements in\nsimulating quantum computers. We then describe a more efficient method that\nstores only the 2n amplitudes of the n qubit state vector |\ψ\⟩\nand optimize its single node performance. In our multi-node implementation, we\nuse a single amplitude communication protocol that maximizes the number of\nqubits able to be simulated and minimizes the ratio of qubits that require\ncommunication to those that do not, and we present an algorithm for efficiently\ndetermining communication pairs among processors. We simulate up to 30 qubits\non a single node and 33 qubits with the state vector partitioned across 64\nnodes. Lastly, we discuss the advantages and disadvantages of our communication\nscheme, propose potential improvements, and describe other optimizations such\nas storing the state vector non-sequentially in memory to map communication\nrequirements to idle qubits in the circuit.\n