Roviero joins Arm AI Partner Program to future-proof AI-based IoT solutions
SAN JOSE, Calif., May 16, 2022 -- Roviero, a Silicon Valley provider of highly programmable and flexible AI compute is joining the Arm® AI Partner Program to bring AI and general-purpose compute as a packaged product to customers in the edge IoT space.
Roviero's CortiCore™ processor architecture is a leading Artificial Intelligence (AI) IP for edge System On Chip (SoC) solutions. CortiCore™ provides best in class power, area and programmability while scaling in performance from less than 1 TOPS to more than 500 TOPS.
Roviero's CortiCore AI IP can be integrated seamlessly with Arm Cortex®-M microcontrollers and high-performance Cortex-A processors, to create edge-based computing devices with integrated machine learning capabilities. There are a plethora of SoC configurations that can be enabled due to configurable options provided by Roviero's IP. Software plays an important role in creating an edge computing device with machine learning capabilities and Roviero's software works seamlessly on Arm-based processors, providing end-to-end integration that can reduce time to market for our end customers.
"In talking to our customers we hear two key concerns in integrating AI IP in their SoC solutions – firstly, efforts needed by their team to integrate and maintain the associated runtime software and secondly future proofing the SoC against new requirements after the SoC has been built. Our team has created a solution based on Arm microcontrollers and associated software that addresses these concerns," said Ravi Setty, Snr Vice President of Engineering at Roviero.
- The CortiCore™ runtime software has been integrated with the Arm host software. With this, customers can instantiate CortiCore™ runtime software with the click of a button. With the instantiation, the application software gets access to the function calls to invoke the AI acceleration and get its response back – simplifying the AI acceleration when run on Arm-based solutions.
- As a CortiCore™ architecture extension Arm microcontrollers are embedded deep into the AI compute. Customers can pick which Arm microcontrollers they need depending on the application space. The runtime software on CortiCore™ seamlessly transfers control to the microcontroller and then back to the AI compute engines. Customers can use their own code on the microcontroller or use Roviero provided functions. The Arm microcontrollers embedded in the IP allow customers to run any custom functions that might be needed in future.
"As more developers look to incorporate AI into IoT devices, simplifying machine learning development is crucial," said Mohamed Awad, vice president of IoT and Embedded at Arm. "Through joining the Arm AI Partner Program, Roviero will help even more developers create the next generation of edge computing devices."
"There is no doubt that AI is the technology needed to create the next generation of solutions and experiences. Roviero's goal is to innovate and work with our customers to implement AI in their SoCs that provides the best in class power, performance, area and programmability. With the high cost of building SoCs, our customers have a dire need to future-proof. By embedding Arm microcontrollers deep into the AI compute, our customers can be confident that they can run any future applications that need AI. We chose Arm because of multiple reasons, including the wide variety of processors for our customers to choose from, ease of integration into our IP and most of all the extensive Arm software ecosystem," said Deepak Mital, CEO at Roviero.
|
Related News
Breaking News
- Jury is out in the Arm vs Qualcomm trial
- Ceva Seeks To Exploit Synergies in Portfolio with Nano NPU
- Synopsys Responds to U.K. Competition and Markets Authority's Phase 1 Announcement Regarding Ansys Acquisition
- Alphawave Semi Scales UCIe™ to 64 Gbps Enabling >20 Tbps/mm Bandwidth Density for Die-to-Die Chiplet Connectivity
- RaiderChip Hardware NPU adds Falcon-3 LLM to its supported AI models
Most Popular
E-mail This Article | Printer-Friendly Page |