International. DENSO Corporation and Quadric.inc have entered into a development license agreement for a Neural Processing Unit (NPU), a semiconductor specializing in artificial intelligence (AI) arithmetic processing.
This agreement will allow DENSO to acquire the license to the intellectual property core of Quadric's Chimera GPNPU, with the aim of co-developing intellectual property innovations (NPUs) for semiconductors in vehicles.
With the rise of intelligent systems such as automated driving and connected technologies, the need to process vast amounts of information from on-board sensors and inter-vehicle and cloud communications in real time has become crucial, according to DENSO. This has increased the importance of in-vehicle SoC performance, which must be able to handle large volumes of information at high speed and with high efficiency.
Both companies have been researching the development of semiconductor IPs (NPUs) using the Chimera GPNPU (General Purpose Neural Processing Unit) for in-vehicle SoCs and have decided to join forces to develop an in-vehicle semiconductor IP (NPU) that combines DENSO's RISC-V-based processor and Quadric's Chimera GPNPU.
DENSO has highlighted that its RISC-V-based processor IP is compliant with the ASIL D ISO26262 standard, making it ideal for automotive applications where safety is a priority. As vehicle intelligence evolves, AI is emerging as a determining factor in the value of cars.
Quadric's Chimera GPNPU features a unique architecture capable of executing matrix, vector, and scalar (control) code operations on a single execution channel. This versatility to handle various computational tasks makes it applicable to multiple AI functions. In addition, the use of the Chimera GPNPU will allow automotive system developers to incorporate unique AI features into their future products, DENSO added.
With this joint development agreement, DENSO and Quadric intend to contribute to the creation of in-vehicle SoCs that respond flexibly to changing AI trends over an extended period, from the product development phase to post-launch, thus driving the creation of AD/ADAS products with high safety standards.
