Arrow Electronics Engineers drive success for AI startup NeuReality
Arrow Electronics played a pivotal role in the creation of the world’s first 7nm Network Addressable Processing Unit (NR1 NAPU), now powering NeuReality’s NR1-S AI Inference Appliance. This technology offers significant advantages in cost and energy savings compared to traditional CPU-centric systems.
The NR1-S, when integrated with AI accelerators in an inference server, can reduce data centre costs by up to 90% and boost energy efficiency by as much as 15 times. It also provides linear scalability without performance degradation, even as more AI accelerators are added, according to NeuReality.
Arrow’s contribution included extensive embedded design expertise, guiding firmware and hardware design, validating power management systems, and debugging the microcontroller (MCU) and platform power flows. They also assisted in selecting the ideal MCU for seamless integration between the PCIe card and server components, ensuring the successful launch of the NR1 NAPU, NR1-S, and integrated NeuReality software – all completed in record time.
The NR1 NAPU is a custom server-on-a-chip that maximises the utilisation of AI accelerators from around 30% to full capacity, significantly enhancing overall output and reducing silicon waste. It handles critical tasks such as network termination, quality of service, and AI data pre-and post-processing, optimising data flow for diverse, high-volume AI pipelines.
NeuReality’s innovative system architecture overcomes the performance bottlenecks of traditional CPU-centric designs, making the NR1-S a game-changer in running AI data pipelines. This advancement addresses the financial and energy efficiency concerns that are central to deploying today’s demanding AI applications.
“Our NAPU addresses the major bottlenecks that restrict performance in today’s AI accelerators, such as power management and transferring data from the network into the AI accelerator, typically a GPU, FPGA or ASIC,” said Eli Bar-Lev, director of hardware at NeuReality. “Arrow’s support with the hardware and firmware for power management and thermal engineering allowed us to focus resources on a complete silicon-to-software AI inference solution which will reduce the AI market barriers for governments and businesses around the world.”
“This exciting project can potentially make cloud and on-premise enterprise AI inferencing more affordable and faster, thereby increasing access to valuable services in healthcare and medical imaging, banking and insurance, and AI-driven customer call centres and virtual assistants,” said Vitali Damasevich, director engineering Eastern Europe and engineering solutions centre EMEA.
