Fujitsu said it has unveiled new technology that to optimise the use of central and graphics processing unites (CPUs and GPUs) by allocating resources in real time.
Fujitisu's technology prioritises processes with high execution efficiency, even when running programs that heavily depend on GPUs.
It is designed to tackle the current shortage of GPUs caused by the growing demand for generative AI and deep learning applications.
By optimising users' existing computing resources, Fujitsu aims to mitigate the shortage of GPUs in various sectors.
Additionally, Fujitsu has also introduced a parallel processing technology that allows for real-time switching between multiple programs in an high performance computing system.
This system, which links multiple computers, eliminates the need to wait for the completion of a running program.
It enables the immediate execution of applications that require large-scale computational resources and real-time performance, such as digital twin and generative AI programs, thus enhancing efficiency and productivity.
Both technologies developed by Fujitsu will be integrated into a future computer workload broker, which currently undergoes development.
This broker will enable AI to automatically select the most suitable resource for a given problem, considering factors such as computation time, accuracy, and cost.
Fujitsu plans to incorporate the CPU/GPU resource optimisation technology into its Fujitsu Kozuchi AI Platform, which allows users to test advanced AI technologies quickly.
The HPC optimisation technology will also be applied to Fujitsu's 40 qubit quantum computer simulator for collaborative computing using a large number of nodes.
-1.jpg&w=100&c=1&s=0)
.png&w=100&c=1&s=0)
