In order to increase its capacity for AI-enabled edge solutions that are power-efficient and installed on field programmable gate arrays (FPGAs), Microchip Technology has purchased Neuronix AI Labs.
Neural network sparsity optimization technique from Neuronix AI Labs allows for high accuracy reduction in computation, size, and power consumption for tasks including object detection, semantic segmentation, and picture categorization.
The industry already looks to Microchip’s mid-range PolarFire® FPGAs and SoCs for their low power consumption, dependability, and security features.
With the acquisition of this Neural network sparsity optimization technique, Microchip will be able to increase the processing power of AI/ML on low- and mid-range FPGAs by a factor of multiple and develop large-scale edge deployments of components intended for use in computer-vision applications on systems with limited resources in terms of money, size, and power.
“The acquisition of Neuronix AI Labs’ technology will enhance our power efficiency for FPGAs and SoCs deployed in intelligent edge systems that utilize AI/ML algorithms,” said Bruce Weyer, corporate vice president of Microchip’s FPGA business unit. “Neuronix technology combined with our VectorBlox™ design flow produces an increase in neural network performance efficiency and delivers outstanding GOPS/watt performance in our low-power PolarFire FPGAs and SoCs. Systems designers will now be able to architect and deploy small-footprint hardware that was previously difficult to build due to size, thermal or power constraints.”
With the acquisition of this technology, industry-standard AI frameworks will enable non-FPGA designers to leverage significant parallel processing capabilities without the need for extensive understanding of FPGA design flow.
Without the requirement for register-transition level (RTL) expertise or in-depth understanding of the underlying FPGA fabric, AI/ML algorithms may be implemented on configurable FPGA logic thanks to the combination of Neuronix AI intellectual property and Microchip’s current compilers and software design kits.
Additionally, it is made to enable real-time CNN upgrades and updates without requiring hardware reprograms.
“Neuronix AI Labs has been laser-focused on producing best-in-class neural network acceleration architectures and algorithms that can transform user expectations of size, power, performance and cost,” said Yaron Raz, CEO of Neuronix AI Labs. “Joining the Microchip team offers us a unique opportunity to scale and align with an FPGA portfolio that has set industry standards for power efficiency.”
To learn more: Microchip’s FPGA and SoC solutions
