GTM-MML4VXJ
Skip to main content

日本語

Japan

  1. Home >
  2. R&D Initiatives >
  3. Innovative Computing for Solving Social Issues

Innovative Computing for Solving Social Issues

Predicting the future of computing and creating a groundbreaking architecture, we will open the way to solve social problems which were difficult to be solved by conventional ICT.

In order to overcome the limits of ICT performance improvement based on semiconductor miniaturization, Fujitsu Labs is tackling the development of a new algorithm and architecture specialized in a certain domain and problem.

The Digital Annealer is our newly developed computing architecture inspired by quantum phenomena, and it has more than 10 thousand times faster computing speed of a conventional computer and can high-speedily solve a large-scale and complex, combinatorial optimization problems which were impossible before. For example, it can solve an intractable problem just in one second, which requires 0.8 billion years for general-purpose computer to solve, such as a travelling salesman problem for visiting 30 cities that needs more than "10 quadrillion x 10 quadrillion" computation, or an investment portfolio optimization problem for 500 brands. We also developed another new technology by combining a supercomputer technology and AI, and increased the simulation speed by 100 times of the conventional method. It means that we can complete a simulation just in one day, which took several months before. In addition to the above mentioned initiatives, we are trying to develop a dedicated processor (DLU) for a large-scale deep learning with power efficiency of more than 10 times of the conventional method, and infrastructure technologies to operate the whole system efficiently, such as power supplying and cooling technologies of ICT equipment, high integration technology, and power control technology for energy saving.

In order to solve social problems which were difficult to be solved by conventional ICT, we are advancing R&D for creating a new computing paradigm and innovative computing architecture.

Related Articles

Back To R&D Initiatives Top