Orion AI brings advanced features such as multicast and broadcast to improve performance and efficiency in artificial intelligence applications. Designed to provide extreme performance—terabits of on-chip bandwidth—its underlying architecture that can support thousands of compute engines. Orion AI provides super-wide data paths with interfaces of up to 1024 bits, higher for internal structures, and can support long bursts of up to 4K Bytes.
Artificial Intelligence (AI) is making its way into many applications including vision, speech, forecasting, robotics and diagnostics. These emerging applications require a whole new level of processing capability and are driving sweeping changes in computational architectures and a dramatic shift in SOC design. New requirements include:
Experience the most powerful and advanced SoC development platform available. Orion AI was designed to provide extreme performance—terabits of on-chip bandwidth. It provides a scalable and reusable SoC interconnect platform that automates IP integration, eliminates deadlocks and improves quality of results.
The underlying architecture can support thousands of compute engines. It provides super-wide data paths with interfaces of up to 1024 bits, higher for internal structures, and can support long bursts of up to 4K Bytes.
First and only interconnect to provide configurable and programmable multicast and broadcast support. Posted write support reduces unnecessary ack messages and increases available bandwidth. The architecture is optimized for peer-to-peer traffic to support the any-to-any communication AI applications need.
Orion AI is powered by NetSpeed’s Turing machine learning engine which uses supervised learning to explore and optimize SoC design and architecture. This is an AI-inside approach that is like having a guru architect on call to provide design advice.
Orion AI comes with NocStudio, NetSpeed’s SoC development platform. Harnessing the power of artificial intelligence, NocStudio delivers the extreme performance next-gen applications need. NocStudio includes tools that enable architects to design, configure and simulate Orion AI as well as evaluate multiple SoC architectures.
Terabits of on-chip bandwidth, 2 GHz+, supports 1000s of cores, super-wide data paths up to 1024 bits, long bursts of up to 4K Bytes.
Advanced low power management techniques to manage energy efficiency. Support for multiple power domains, voltage domains through industry standard power control interface to support management of power state transitions.
NetSpeed’s Turing machine learning engine uses supervised learning to explore and optimize SoC design and architecture. It's like having a guru architect on call to provide design advice.
Robust, configurable end-to-end QoS schemes: Bandwidth allocation, strict non-blocking priority levels, rate limiters.
Up to 32 virtual networks support in the NoC; Virtual channels are allocated to ensure QoS, deadlock avoidance and bandwidth intent satisfaction.
Posted write support reduces unnecessary ack messages to increase available bandwidth
First and only interconnect to provide configurable and programmable multicast and broadcast support
Architecture optimized to perform rapid peer to peer data exchange rapidly and efficiently
Integrated performance simulator that empowers architects to understand the tradeoffs among power, performance, and area with real-time feedback.
Integrated architectural exploration platform simulates various application use-cases and industry benchmarks.
Optimize either memory sub-systems and complete SoC systems using end-to-end bandwidth, latency and power tradeoff metrics.
Powerful data analytics tools that provide insights on the performance and power bottlenecks in the system.
Perform power, performance,area tradeoff analysis with realtime feedback on SoC bandwidth, latency and power.
See what our customers are saying about Orion AI
"NetSpeed is the unquestionable leader in providing a high performance interconnect solution for the AI market segment. The reason we choose NetSpeed is the scalability of the IP and the ability to configure sophisticated topologies for high performance required by AI applications with minimal effort. Hence, we are choosing NetSpeed’s IP with its advanced AI extensions for our next-gen SoC designs."
"Our chip technology needs to be very flexible and scalable to accommodate a vast range of power, performance and area requirements. NetSpeed's interconnect IP enables us to rapidly adapt our platform to meet the specific requirements of a wide variety of embedded AI applications. We chose NetSpeed’s technology because of its higher performance, full coherency support, configurability, and scalability. We believe NetSpeed IP is a great on-chip network solution that fully addresses the needs of next-generation SoCs targeting AI applications."
"Esperanto is putting thousands of processors and accelerators on a single chip, and our challenge was how to interconnect them. NetSpeed provided a compelling solution for interconnecting our cores with high performance and by licensing their IP we can reduce our time to market as well."
Whether you are a hardware architect looking for cache coherency solutions to realize your ideas into silicon or a SoC designer looking for a high performance and efficient interconnect or a safety manager looking for solutions to build resilient, fault-tolerant systems—NetSpeed has a solution for you.
Crafted @ Lollypop.biz