You may think that NVIDIA doesn’t have any footsteps in the mobile department, but that seems to been done with great ambition to face a completely new set of challenges. Now this is the case in its efforts in the field of autonomous cars or self driving cars, where we have already started to witness their products starting to prove very interesting for the automotive industry.
The efforts of NVIDIA continues with the presentation of Xavier, AI computer System on Chip (SoC) which integrates a new GPU (Graphics Processing Unit) – based architecture named “Volta”.
Volta also has an 8-core CPU(Central Processing Unit) which is aimed in acceleration of visual recognition environment of the car. NVIDIA ensures that its new supercomputer to the car can achieve 20 TOPS(Trillion Operations Per Second) performance and do well with a consumption of just 20Watts of power. Xavier is intended to be compliant with crucial automotive standards, like the ISO 26262 functional safety specification.
The brain of the car (and truck and bus) could be the Xavier soon, as announced by the CEO of NVIDIA during the launch of Xavier. The combination of this could be the key to the success of this SoC which targets to become the replacement to most of today’s SoC, being more powerful and especially more efficient when it comes to providing the required magnitude of computing power for an autonomous car with sufficient hardware resources.
The NVIDIA’s new SoC in Xavier is intended for the future autonomous vehicles which has 7 billion transistors and is manufactured using a cutting edge process of 16 nm FinFET lithographic. If you want to learn more about what an SoC is, you could check out our previous post dedicated to SoCs.
In the official announcement NVIDIA points out that these processors want to fight with all types of vehicles, and mentions not only cars, but also the trucks, buses and taxis in the case of automation with xavier. Xavier still need some time to get launched, however, it will be ready for manufacturers and research developers for use from the fourth quarter of 2017.
If interested, you could check out the entire NVIDIA Keynote at the GPU Conference 2016 right here.. uploaded by ExpoVistaTV
Image credits NVIDIA