Machine learning for safety-critical functions? Kalray says yes: Page 2 of 2

December 12, 2018 //By Christoph Hammerschmidt
Machine learning for safety-critical functions? Kalray says yes
With a massive parallel processing architecture (up to 288 cores), startup company Kalray (Grenoble, France) claims superior performance for compute-intensive real-time tasks, in particular for Artificial Intelligence applications. In an exclusive interview with eeNews Europe, Stéphane Cordova, Vice President of Kalray’s Embedded Business unit, explains why these processors are beneficial in automotive environments.

Competing architectures rely on external memory – for every memory access they need to go outside the chip.Thus, for them, memory access takes much more time. In addition, we have extra coprocessors dedicated to machine learning. This combination yields very high performance and very low power consumption at the same time. Another factor is that our competitors have a big legacy – installed base, ecosystem, and existing software - that prevents them from quick changes. Rewriting these applications would be a nightmare.

eeNews Europe: As a startup-company, you may have the advantage of being more agile and flexible, but the products of your competitors are probably more mature.

Cordova: We have ten years of experience; our products are now in the third generation. Our products are mature enough to go into cars within the next years. Our processor is based on our own patented core design. In terms of software, they are programmed in standard C++; existing applications can be ported to our architecture almost for free.

Related articles:

Kalray: MPPA processors for autonomous driving

France, Renault rally round Kalray

Loss-making Kalray raises funds via IPO

Kalray's Coolidge processor adds deep learning acceleration

 


Vous êtes certain ?

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

Vous allez être rediriger vers Google.