Argonne scientists, led by researchers Yuri Alexeev (CPS), Hal Finkel (LCF) and Martin Suchara (MCS), contributed to three quantum computing workshop papers which will be presented at the International Conference for High Performance Computing, Networking, Storage, and Analysis in Dallas next week.
The six-day conference, known as SC18, is now in its 30th year and kicks off Nov. 11.
The first paper, titled “Community Detection Across Emerging Quantum Architectures,” introduces a new algorithm to boost the efficiency of quantum computers.
“It was applied to clustering real-world networks of up to 2,000 nodes using only 16 qubit IBM Q and D-Wave 2000Q quantum computers,” said Alexeev, a principal project specialist at the Argonne Leadership Computing Facility.
Alexeev has a background in quantum chemistry, molecular dynamics and high performance computing (HPC). He has been working on a number of quantum computing and networking projects since 2017. Ruslan Shaydulin, a Ph.D. student at Clemson University and a summer intern at Argonne, conducted the study under Alexeev’s direction to use quantum computers to solve research problems.
The next paper, “Memory-Efficient Quantum Circuit Simulation by Using Lossy Data Compression,” introduces an algorithm to compress data in a part Argonne-developed quantum simulator so that it could run large-scale quantum simulations.
Xin-Chuan, a Ph.D. student in the Department of Computer Science (Systems Group) at the University of Chicago, authored the second paper — also under Alexeev’s and Finkel’s guidance.
The final paper is the result of Alexeev’s close collaboration with computational scientists Finkel and Suchara, whose research focuses on quantum communication and networking, quantum error correction, quantum simulations and making quantum computation more efficient and reliable.
Their paper, “Hybrid Quantum-Classical Computing Architectures,” describes how supercomputing can help modern small quantum computers solve large problems.
“It is achieved by breaking large quantum circuits into smaller sub-circuits that are computed separately, either using a quantum computer or a quantum simulator running on a supercomputer,” Alexeev said.
SC18 organizers say their upcoming event provides the largest and most diverse technical program of any HPC conference.
“Our technical program is designed to share the latest research and best practices in algorithms, architectures, benchmarking, large-scale system deployments, system performance, applications, resiliency and the convergence of big data and extreme computing,” said SC18 Conference Chair Ralph A. McEldowney in a note to attendees. “The SC18 technical program is truly six months of education and training jammed into six days.”
SC18 will include more than 386 exhibitors showcasing “all aspects of the HPC ecosystem such as computing, networking, storage, analytics and applications,” McEldowney said.
By Jo Napolitano