The architecture boasts over 1000 cores, each equipped with its own memory, which the company claims will 'revolutionise' traditional Von Neumann computing paradigms and advance ultra-low power in-memory computing.
This milestone came about through Green Mountain’s work in support of Non-Von LLC, and their patented innovative chip design.
Green Mountain Semiconductor has been a frontrunner in a number of AI-related circuit design initiatives, particularly in the development of ultra-low power processing in-memory neural networks.
The newly designed ASIC computing architecture looks to address the inherent limitations of traditional Von Neumann architectures, which often face challenges in efficiently exchanging data between memories and central processing units (CPUs). By providing each core with its dedicated memory, Green Mountain Semiconductor's solution enhances data access and processing speed while significantly reducing power consumption.
"This ground-breaking ASIC computing architecture marks a significant departure from traditional computing paradigms," said Wolfgang Hokenmaier, President at Green Mountain Semiconductor. "Our approach not only addresses critical limitations but also aligns with our vision of revolutionising data processing for various applications, including high-complexity neuromorphic computing and low-power devices such as those used in medical and IoT applications."