GSI Technology, Inc., has announced its new APU-based processor delivers significant energy savings when powering artificial intelligence (AI) search applications for data centers as well as improving customers’ efforts to improve the environment.
Lee-Lean Shu, chairman, and CEO, GSI Technology, said, “From the migration of compute to the cloud to the increased need for on-prem business hardware, the race is on to acquire energy-efficient hardware and lower the power bill. Solving this problem is one of the strategic opportunities for our technology. Our new APU processor offers significant advantages in power consumption, enabling enterprise customers to reach their sustainability goals and do good for the planet. GSI has shown in comparison for a large area SAR image in one second at high resolution that the Gemini-I APU used on average 93% less power than CPU or GPU systems, in a portable solution versus requiring a room of server cabinets.”
Increasing energy consumption has been hidden behind closed doors due to the concentration of computing technologies in enterprise data centers. Although these decades-old, compute-centric, and I/O-bound architectures are updated to take advantage of the latest process technology, there is a significant impact on user costs, energy consumption, and the environment. CPUs, FPGAs, and GPUs that utilize traditional compute methodologies are not helping when they pack more capacity and more cores or just increase density.
Data-intensive and AI-centric workloads will likely exacerbate this problem. The design of future AI-centric data centers must continue to utilize new hardware and chipsets specifically designed for power efficiency. GSI’s new technology, which saves enormous amounts of energy, helps to solve this critical issue. APU-based GSI servers have a much smaller energy footprint and provide dramatic power savings compared to traditional server systems.