NeuReality, a semiconductor company located in Israel that is developing the next generation of AI-centred computing system architecture, has partnered with IBM to produce the next generation of high-performance AI inference platforms. These systems, according to both companies, will yield significant cost and power savings for deep learning use cases. This follows NeuReality’s public debut in February when the company raised $8 million in a seed round to accelerate AI workloads at scale.
According to the official announcement, this alliance will enable important industries including banking, insurance, healthcare, manufacturing, and smart cities to implement computer vision, recommendation systems, Natural Language Processing, and other AI use cases. They also say that the partnership would hasten deployments in today’s rapidly expanding AI use cases, which are currently available in public and private cloud data centers.
AI Inference is an emerging buzzword in the artificial intelligence domain. It is the aspect of AI where neural networks are really employed in real-world applications and generate results. While high-performance AI inference systems are becoming a rising area of attention for businesses, they also promise to minimize the cost and power consumption of deep learning.
The agreement covers NR1, NeuReality’s first AI-centric architecture Server-on-a-Chip ASIC implementation. The NR1 prototype platform is built on NeuReality’s first-generation FPGA-based NR1-P prototype platform, which was unveiled earlier this year in May. According to the company, with features like native AI-over-Fabric networking, full AI pipeline offload, and hardware-based AI hypervisor, NR1-P can eliminate system bottlenecks in today’s solutions. In addition, it will also lower the cost and power consumption of inference systems and services. Prior to the release of the NR1 production platform next year, the NR1-P platform will offer software integration and system-level validation.
As part of the newly formed deal, IBM will become a NeuReality design partner, working on product requirements for the NR1 chip, system, and SDK, which will be incorporated in the architecture’s next edition. The two businesses will jointly examine NeuReality’s solutions for usage in IBM’s Hybrid Cloud, covering AI use cases, system flows, virtualization, networking, security, and other areas.
“Having the NR1-P FPGA platform available today allows us to develop IBM’s requirements and test them before the NR1 server-on-a-chip’s tapeout,” said NeuReality CEO Moshe Tanach. “Being able to develop, test and optimize complex data center distributed features, such as Kubernetes, networking and security before production is the only way to deliver high quality to our customers.”
The basic NR1-P architecture is based on a 4U server chassis with 16 Xilinx Versal Adaptive Compute Acceleration Platform (ACAP) cards. According to NeuReality, the NR1-P platform, which is intended for usage in cloud data centers and edge nodes, is now being shown to clients and partners and will be followed by further implementations. Customers may incorporate it in orchestrated data centers and other facilities and witness it in action on the new prototype platform, which is being used to test the technology.
IBM opened the IBM Research AI Hardware Center in Albany, New York, in February 2019 to build a worldwide research hub for developing next-generation AI hardware with multiple technological partners and expanding the company’s cooperative nanotechnology research activities. NeuReality will become the IBM Research AI Hardware Center’s first start-up semiconductor product member and a licensee of the Center’s low-precision high-performance Digital AI Cores as a result of this cooperation.
Dr. Mukesh Khare, Vice President of Hybrid Cloud research at IBM Research, said, “In light of IBM’s vision to deliver the most advanced Hybrid Cloud and AI systems and services to our clients, teaming up with NeuReality, which brings a disruptive AI-centric approach to the table, is the type of industry collaboration we are looking for.” Mukesh added that the partnership with NeuReality is expected to drive a more streamlined and accessible AI infrastructure, which has the potential to enhance people’s lives.