Anticipating future computing requirements
Computing performance is already being substantively enhanced by a wide range of techniques such as parallel processing, accelerators, and high-speed calculation algorithms. These achieve performance improvements without relying on semiconductor microfabrication, but do require expert intervention in order to optimize their operation.
Fujitsu Laboratories is adopting a different approach, building on the proven platform of Fujitsu’s strengths in supercomputers, domain-specific accelerator technologies, new data processing architectures, and edge computing technologies. Our aim is to achieve extraordinarily high computing performance through the application of parallel processing technologies on a massive scale. In addition, we are working to create an environment that ensures every level of software programmer can use the technology to achieve significant performance improvements.
Our R&D initiatives spotlight problem-solving at every layer of a user’s ICT system. For next generation computing, we have developed DLU: our Deep Learning domain-specific accelerator and parallel technology. Our intelligent data store technology sets new levels of performance and reduces storage costs via its hierarchical data storage characteristics based on data content. Our real-time edge computing utilizes edge-side resources to set new standards.