November 15, 2022
Fujitsu is attempting to make hybrid computing more accessible by developing technology to optimize workload selection across quantum and high-performance computing (HPC) processors for customers.
Quantum computers need to operate alongside classical computers to deliver solutions to real-world problems, a combination known as hybrid computing. Working out which algorithms that form a solution should run on each platform has required complex technical knowledge.
The new AI-based software automatically selects the right combination of platforms to process problems in an optimal way using parameters including calculation time, calculation accuracy and costs.
Fujitsu’s technology uses the company’s 39-qubit quantum simulator running on its supercomputer, which uses the same CPU as its Fugaku petascale supercomputer at the Riken Center for Computational Science in Kobe, Japan.
According to Fujitsu, the technology allows users without specialized knowledge to use quantum simulators and HPC technology to develop effective solutions to real-world problems. The initial use case will be combining and selecting the optimum calculation method for customers’ quantum chemical calculation problems.
Fujitsu intends to develop the solution into what it calls computer workload broker software. This would use artificial intelligence to automatically select the appropriate resources from a range of technologies including large-scale cloud-based HPC, quantum computing, Fujitsu’s quantum-inspired Digital Annealer and quantum simulators to solve customers’ complex problems.
The computing workload broker can support the automation and optimization of problem-solving across multiple hardware devices and the optimization of calculation distribution for different platforms. This would enable high-speed computing tailored to specific use cases.