Fujitsu develops quantum/HPC hybrid computing technology

- Advertisment -

Today, Fujitsu announced the creation of quantum/HPC hybrid computing technology to enhance customer workload selection.

In a statement, Fujitsu said the new AI-based software automatically chooses from many next-generation computing platforms to provide the best answer to clients’ problems based on factors like calculation speed, calculation accuracy, and cost. It acts as a forerunner to future computer workload broker technology.

The new technology combines and chooses the best calculation approach for customers’ quantum chemical calculation problems using Fujitsu’s 39-qubit quantum simulator and the “FUJITSU Supercomputer PRIMEHPC FX 700,” which is powered by the same A64FX CPU as supercomputer Fugaku. Notably, the technique enables users without specialised understanding to use HPC and quantum simulators to effectively solve real-world issues.

World-leading advanced computing technologies for solutions to 21st-century problems

Fujitsu has a long history of developing and providing world-leading computing technologies, as evidenced by the development and commercialisation of the supercomputer Fugaku, next-generation quantum computing, and technologies that bridge the conventional and quantum divide, such as the Digital Annealer and quantum simulators.

Fujitsu unveiled the world’s fastest quantum simulator in March 2022 to support the practical creation of novel quantum algorithms, kicking out joint research in materials science with Fujifilm Corporation and Tokyo Electron Ltd.

Easy access to next-generation computing technologies for non-computing experts

Fujitsu is currently developing a computing workload broker—a software technology that utilises AI to automatically choose the right resources from advanced computing technologies and solutions, such as large-scale cloud-based HPC, quantum computing, Digital Annealer, and quantum simulators, to solve customers’ complex problems.

Fujitsu envisages the future computing workload broker as facilitating both the automation and optimisation of calculation processing across numerous hardware devices and the optimisation of calculation distribution across multiple platforms. In this approach, the computing workload broker will enable high-speed calculations and solutions tailored to the demands of individual consumers.

Outline quantum/HOC hybrid computing technology

Fujitsu has recently created a quantum/HPC hybrid calculation technique for tackling quantum chemical challenges, enabling calculation-based clarification of the features of materials used in drug discovery and novel material development. The new technology, which serves as a predecessor to the computing workload broker, enables high-accuracy calculations at high speed by automatically and optimally merging two types of computers: a quantum simulator and HPC. The following are the new technology’s characteristics:

  • Quantum/algorithm discrimination technology – Quantum chemical algorithms necessitate repeated calculations until a very accurate result is obtained. As the distance between atoms changes, existing technologies face a challenging task in identifying whether quantum or HPC algorithms offer the best answer to a problem. To that purpose, Fujitsu created a solution that analyses how an algorithm converges on molecules to discover the most accurate method for an issue. For problems where classical algorithms fail to deliver satisfactory results, the technology detects a specific pattern in the convergence state until the algorithm calculates a solution, allowing the optimum algorithm to be determined by performing experimental pre-processing on the problem using HPC algorithms.
  • Computation time estimation technology – Accurate assessment of the convergence of diverse molecule structures was a difficult task in quantum chemical computations, and it was difficult to predict the time and expense required to get extremely accurate solutions ahead of time. To address this issue, Fujitsu created an AI model capable of estimating calculation quantities, costs, and time in advance by analysing the relationship between the molecular structure, the iterative calculation of the algorithm, and the calculation time.
  • Optimal control technology based on time, cost, and accuracy – Fujitsu created a technique that optimises the efficiency of quantum chemical calculations to enable customers to do computations at the lowest possible cost and in the shortest possible time. The technology considers the calculation time and cost estimated by the quantum/HPC algorithm discrimination technology, the computation time estimation technology, and the utilisation of calculation resources. As a result, the technology enables users to solve quantum chemical calculation issues in a way that best matches their demands without having to select between different computer resources.

Moving forward, Fujitsu will test and refine this technology, with the goal of building a new computing workload broker technology in the field of quantum chemical calculations by fiscal 2023. Fujitsu will continue to develop computing platforms that anyone can use without requiring expertise by adding services to Fujitsu Computing as a Service (hereinafter “CaaS”), Fujitsu’s service portfolio that provides customers with access to world-leading computing technologies via the public cloud.

LATEST NEWS

- Advertisment -