An Anonymous Coward writes:
Nvidia has just announced their new supercomputer that will be used to train AI networks for self-driving cars. See https://www.autonomousvehicleinternational.com/news/computing/nvidia-supercomputer-to-handle-ai-data-from-autonomous-vehicles.html
Full text of the press release is below. There is a photo of the installation at the link:
Computing technology developer Nvidia has built the world’s 22nd fastest supercomputer – DGX SuperPOD – to provide the AI infrastructure needed to meet the demands of the company’s autonomous vehicle deployment program.
DGX SuperPOD was built in just three weeks using 96 Nvidia DGX-2H supercomputers and Mellanox interconnect technology. Delivering 9.4 petaflops of processing capability, it has the power needed for training the vast number of deep neural networks required for safe self-driving vehicles.
A single data-collection vehicle generates 1TB of data per hour. Multiply that by years of driving over an entire fleet, and you quickly get to petabytes of data. That data is used to train algorithms on the rules of the road — and to find potential failures in the deep neural networks operating in the vehicle, which are then re-trained in a continuous loop.
“Few AI challenges are as demanding as training autonomous vehicles, which requires retraining neural networks tens of thousands of times to meet extreme accuracy needs,” said Clement Farabet, vice president of AI infrastructure at Nvidia. “There’s no substitute for massive processing capability like that of the DGX SuperPOD.”
Powered by 1,536 Nvidia V100 Tensor Core GPUs interconnected with Nvidia NVSwitch and Mellanox network fabric, the DGX SuperPOD hardware and software platform takes less than two minutes to train ResNet-50. When this AI model came out in 2015, it took 25 days to train on the then state-of-the-art system, a single Nvidia K80 GPU. DGX SuperPOD delivers results that are 18,000 times faster.
While other TOP500 systems with similar performance levels are built from thousands of servers, DGX SuperPOD takes a fraction of the space and is roughly 400 times smaller than its ranked neighbors.
Will this be enough computrons? It seems every time another announcement is made in this field, it includes yet more compute power to train AIs on ever larger data sets. From what your AC has seen, there is still a good way to go before these network attached cars can match the competence of a good driver (not impaired, not distracted)--which might be, imo, one reasonable target before wide release of the technology.
Alternatively, will someone come up with better AI algorithms (more like people?) that will vastly change/reduce the amount of training required?
Related Stories
Nvidia has announced its next chip for self-driving cars years in advance:
First outlined as part of NVIDIA's DRIVE roadmap at GTC 2018, NVIDIA CEO Jensen Huang took the stage at GTC China this morning to properly introduce the chip that will be powering the next generation of the DRIVE platform. Officially dubbed the NVIDIA DRIVE AGX Orin, the new chip will eventually succeed NVIDIA's currently shipping Xavier SoC, which has been available for about the last year now. In fact, as has been the case with previous NVIDIA DRIVE unveils, NVIDIA is announcing the chip well in advance: the company isn't expecting the chip to be fully ready for automakers until 2022.
What lies beneath Orin then is a lot of hardware, with NVIDIA going into some high-level details on certain parts, but skimming over others. Overall, Orin is a 17 billion transistor chip, almost double the transistor count of Xavier and continuing the trend of very large, very powerful automotive SoCs. NVIDIA is not disclosing the manufacturing process being used at this time, but given their timeframe, some sort of 7nm or 5nm process (or derivative) is pretty much a given. And NVIDIA will definitely need a smaller manufacturing process – to put things in comparison, the company's top-end Turing GPU, TU102, takes up 754mm2 for 18.6B transistors, so Orin will pack in almost as many transistors as one of NVIDIA's best GPUs today.
[...] All told, NVIDIA expects Orin to deliver 7x the 30 INT8 TOPS performance of Xavier, with the combination of the GPU and DLA pushing 200 TOPS. It goes without saying that NVIDIA is still heavily invested in neural networks as the solution to self-driving systems, so they are similarly heavily investing in hardware to execute those neural nets.
[...] Finally, while NVIDIA hasn't disclosed any official figures for power consumption, it's clear that overall power usage is going up relative to Xavier. While Orin is expected to be 7x faster than Xavier, NVIDIA is only claiming it's 3x as power efficient. Assuming NVIDIA is basing all of this on INT8 TOPS as they usually do, then the 1 TOPS/Watt Xavier would be replaced by the 3 TOPS/Watt Orin, putting the 200 TOPS chip at around 65-70 Watts. Which is admittedly still fairly low for a single chip at a company that sells 400 Watt GPUs, but it could add up if NVIDIA builds another multi-processor board like the DRIVE Pegasus.
The design will include 12 ARM "Hercules" (Cortex-A78) cores rather than Nvidia-designed custom ARM cores.
Also at Wccftech.
Related: Nvidia Demos a Car Computer Trained with "Deep Learning"
Nvidia Announces Jetson Nano Single-Board Computer
Nvidia Supercomputer to Crunch Autonomous Vehicles Data
(Score: 2) by Pslytely Psycho on Saturday June 22 2019, @07:55AM (1 child)
“Listen,” said Ford, who was still engrossed in the sales brochure, “they make a big thing of the ship's cybernetics. A new generation of Sirius Cybernetics Corporation robots and computers, with the new GPP feature.”
“GPP feature?” said Arthur. “What's that?”
“Oh, it says Genuine People Personalities.”
“Oh,” said Arthur, “sounds ghastly.”
A voice behind them said, “It is.” The voice was low and hopeless and accompanied by a slight clanking sound. They span round and saw an abject steel man standing hunched in the doorway...
“Ghastly,” continued Marvin, “it all is. Absolutely ghastly. Just don't even talk about it. Look at this door,” he said, stepping through it. The irony circuits cut into his voice modulator as he mimicked the style of the sales brochure. “All the doors in this spaceship have a cheerful and sunny disposition. It is their pleasure to open for you, and their satisfaction to close again with the knowledge of a job well done.”
As the door closed behind them it became apparent that it did indeed have a satisfied sigh-like quality to it. “Hummmmmmmyummmmmmm ah!” it said...
"Thank you the marketing division of the Sirius Cybernetics Corporation," said Marvin, and trudged desolately up the gleaming curved corridor that stretched out before them. "Let's build robots with Genuine People Personalities," they said. So they tried it out with me. I'm a personality prototype. You can tell, can't you?"
Alex Jones lawyer inspires new TV series: CSI Moron Division.
(Score: 0) by Anonymous Coward on Saturday June 22 2019, @02:05PM
The HHGTTG dystopia is still a few (light) years off...
Closer to home, I saw something recently about relatively small sounding tweaks to the organization of back propagating neural networks. The idea was to add a touch more "intelligence" to the pattern recognition and the claim was it would reduce the size of training sets dramatically. Sorry, can't find the reference now, might have been in Technology Review?
(Score: 0) by Anonymous Coward on Saturday June 22 2019, @01:00PM (2 children)
But who wants to deal with nvidia and all their lies and backstabbing partners/investors to use this stuff?
(Score: 1, Insightful) by Anonymous Coward on Saturday June 22 2019, @01:59PM
There are still some car companies that don't have their own autonomous development effort, and are not partnered with one of the large tech company efforts -- that's who nVidia is trying to attract. I think Intel may be doing something similar?
A related question, would it be worse to deal with nVidia...or with Uber? Plenty of bad corporate actors are trying to shoehorn themselves into this new business.
(Score: 1, Insightful) by Anonymous Coward on Saturday June 22 2019, @08:49PM