Fujitsu’s latest computational hardware could revolutionise the way doctors handle patient data.
Quantum computers are often held aloft as the future of computing. The trouble is, nobody has built one yet.
With the announcement of Fujitsu’s new Digital Annealer Unit (DAU) at this year’s Fujitsu Forum in Tokyo, the Japanese company believes it’s about to usher in an entirely new era of high-performance computing without the pitfalls associated with quantum computing. Fujitsu see the digital annealing technology as a stopgap between current computing methods and the practicalities of quantum computing.
Working with both the University of Toronto and Canadian quantum computing firm 1QBIT, Fujitsu says it has proven digital annealing can slice complex processing times down from years to mere hours, minutes or even seconds. The company claims this could have an enormous impact on everything from economic trading to fighting cancer.
What is digital annealing?
The easiest way to visualise the process is by trying to fill a box with irregular shapes. Current processing methods rely upon an iterative approach of placing shapes into a box until they all fit. Digital annealing randomly places the shapes and then shifts everything around until it all neatly fits. By moving towards this way of processing, Fujitsu discovered they could seriously cut down on processing times around huge datasets.
To make digital annealing a possibility, Fujitsu developed the DAU hardware solution while working with 1QBIT to develop the software layer researchers and developers need to make use of the technology. Speaking on stage at Fujitsu Forum, 1QBIT CEO Andrew Fursman claimed that the company’s “digital annealer surpassed any other hardware the team had used up to that point.” He even stated that it’s the best hardware for developing and using the software 1QBIT is building.
Why is digital annealing such a big deal?
During DAU pilot tests with the University of Toronto, the processor smashed through complex route planning possibilities, analysed molecular data and looked at effective cancer treatment methods by analysing patient data. One example given by Fujitsu Laboratories head Shigerus Sasaki was DAU’s ability to analyse patient data so oncologists could deliver real-time treatment to cancer patients while minimising damage to the surrounding tissue.
In short, DAU is a number-crunching powerhouse, and the possibilities developed during its pilot phase only scratch the surface of what the chip could do.
“We believe that the applications [for digital annealing] won’t be limited to just research and development activities,” Fujitsu president Tatsuya Tanaka explained. “For instance, the combinatorial optimisation problems that were processed in the past took up so much development time. By using digital annealing it frees up time and allows for the development of new services and opportunities and possibly even new discoveries.”
Sasaki believes the market potential for DAU is incredible too, with the medical industry only one of many areas the process could revolutionise.
“The business of selling equity and stocks could benefit [from digital annealing] as it currently requires vast computational power to create accurate reports so they summarise data instead to make processing it easier,” he explained. “With more detailed information, more precise buy and sell calculations could be performed, meaning that there’s virtually no loss reported and that benefits economic activity.”
Fujitsu isn’t the only player in the digital annealing space either. Microsoft has made hints that it’s looking towards a similar solution, albeit one that works via optical computing and is less focused on high computational performance like Fujitsu’s solution.
The first generation chip is available now, with new versions planned for the future to open up a scalable infrastructure to those who find themselves needing even more power than DAU currently offers.