Neural Network processor boosts performance of radar, lidar, vision applications: Page 2 of 2

May 02, 2017 //By Christoph Hammerschmidt
Neural Network processor boosts performance of radar, lidar, vision applications
It is generally accepted that automated driving requires huge computing capacities. What remains subject to discussion however is which processor architecture is best suited to run the object identification and sensor fusion algorithms that enable computers to drive vehicles. Now Cadence has, through its Tensilica division, introduced a processor based on a Digital Signal Processor (DSP) architecture. The Tensilica Vision C5 is designed to run all neural network layers of an AI engine.
Tensilica Vision C5
Wide data path, SIMD execution are the secrets
of the Vision C5 performance

The power efficiency of the C5 is about one order of magnitude better than that of comparable GPU-based solutions, Cadence claims. It also offers an edge over neural network hardware accelerators: Since it is reprogrammable (in contrast to specialized, hardware-based accelerators) it is more flexible and future-proof. And since its functionality is implemented entirely in software with the development tools being widely available, the development is much easier, says Desai. Neural Network libraries will be available once the final version of the C5 will hit the markets. Besides automotive applications which certainly will represent the largest market, Cadence also plans to sell it to manufacturers of drones and security systems or any other kind of applications where neural-network based recognition algorithms are used to perform object detection and recognition.

 

Related articles:

Cadence’s Tensilica Fusion G3 DSP features efficient C programming

Dream Chip shows ADAS processor on FDSOI

Embedded ADAS Algorithm Optimization with High-Performance DSP IP and CV Software Library

Renesas tackles autonomous driving with scalable platform

Image signal processor makes cars see

 


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.