Efficient Technologies has announced a strategic partnership with GlobalFoundries (GF) to bring to market a revolutionary high-performance computer processor. This new processor is reportedly up to 166 times more energy-efficient than the current industry-standard embedded CPUs. The groundbreaking project, which is already in collaboration with select customers, aims for early access and customer sampling by summer 2025, marking a significant milestone in computing.
Innovating Energy Efficiency
The new processor aims to redefine computing by eliminating restrictive energy limitations. It combines Efficient Technologies’ novel architecture and technology with GF’s U.S.-based manufacturing capabilities, global reach, and market expertise. This collaboration is set to enable a quantum leap in edge device capabilities and battery lifetime.
Current general-purpose processors are often over-designed for generality, leading to considerable energy consumption due to unnecessary internal data movement and instruction control overheads. By contrast, the Efficient fabric processor architecture provides reconfigurable hardware at compile time, achieving up to 99% lower DC power consumption without compromising performance. This paradigm shift sets a new benchmark in ultra-low energy and high-performance processors.
Applications and Future Prospects
The partnership between Efficient and GF is expected to unlock computing power for smarter applications, enabling scenarios previously impossible due to energy constraints. When combined with GF’s 22FDX platform’s MRAM and adaptive body biasing (ABB) capabilities, Efficient’s new processor can significantly reduce power consumption, facilitating advanced capabilities in devices worldwide.
The Efficient fabric processor is versatile, supporting various applications such as machine learning-enabled extreme-edge machine vision, continuous audio intelligence, and sensory and signal intelligence. It also allows developers to utilize their code, supporting popular embedded languages like C and TensorFlow Lite (TFLite) at launch, with plans to expand language support within the first year. This flexibility will enable the creation of sophisticated intelligence applications incorporating artificial intelligence, machine learning, signal processing, data analytics, and other general-purpose data processing tasks.
Conclusion
The strategic partnership between Efficient and GlobalFoundries marks a significant advancement in the field of energy-efficient computing. By addressing the constraints of power consumption, this collaboration is set to pave the way for the deployment of powerful, energy-efficient edge devices, propelling innovation and enabling previously unimaginable use cases.
