Nvidia launches self-driving simulation system, Kubernetes support

Posted March 29, 2018

Nvidia CEO Jensen Huang said at the company's annual GPU Conference in San Jose that the cloud-based system, Nvidia DRIVE Constellation, was based on two different servers, with no need for a physical vehicle. The partnership will also bring higher levels of performance to deep learning developers as they leverage Arm's flexibility and scalability.

While NVIDIA didn't announce any future GPU architectures at this year's GTC, the company did try to make the case that its combo of hardware and software improvements is advancing performance at a breakneck pace.

Chipmaker Nvidia said it has suspended global self-driving vehicle testing. Constellation offers a safer way to test self-driving cars, to prevent accidents like the one involving Uber in Arizona last week. The NVSwitch takes the innovations that came from NVIDIA NVLink and extends them further, aided the build of even more advanced and complex systems. "We view the automotive pause as temporary and continue to view Nvidia positively for gaming, virtual reality, auto and artificial intelligence".

Like any machine learning algorithm, NVIDIA's self-driving technology will only get better with every simulation. To begin with, the system houses an impressive array of 16 Tesla V100 GPUs spread across two separate GPU boards. "We don't know exactly what happened", he said. Having multiple servers helps with this.

The second server features Nvidia Drive Pegasus AI auto computer that runs the autonomous vehicle software stack.

Uber's self-driving fleet has been grounded and now Nvidia's is, too. This data is processed just like it is actually coming from the sensors of a vehicle on the road-at speeds of 30 times per second.

Centre issues notice to Cambridge Analytica
It said the information was deleted at Facebook's request, and that it has been in touch with the U.K. Cambridge Analytica has said that "it is willing to cooperate with the ICO in its investigation".

Driving commands from Drive Pegasus will be fed back to the simulator, finishing the digital feedback loop. The second contains a powerful NVIDIA DRIVE Pegasus™ AI auto computer that runs the complete autonomous vehicle software stack and processes the simulated data as if it were coming from the sensors of a auto driving on the road.

Drive Constellation allows automakers and other developers to validate and strengthen the technology through billions of driving miles that allow for testing in hard scenarios. Drive Constellation can run thousands of virtual worlds, each while running thousands of scenarios in order to collect more data.

One major advantage of DRIVE Sim software is photorealistic images. It can simulate different weather patterns including rainstorms and snowstorms, blinding glare at different times of the day, limited vision during night times and all different types of road surfaces and terrain.

"This means you can easily test rare and hard conditions: rainstorms, snowstorms, and sharp glare at different times of the day and night, with different road surfaces and surroundings", the company says.

"Autonomous vehicles need to be developed with a system that covers training to testing to driving." siad Luca De Ambroggi, research and analyst at IHS Markit. "Many of these advances stand on NVIDIA's deep learning platform, which has quickly become the world's standard".

Drive Constellation will begin rolling out to early access partners in the third quarter. Nvidia is working with 370 companies that are all developing autonomous vehicles in some way, including Uber.