Nvidia’s 2k teraflop super chip for self-driving and assisted driving
On Tuesday 20 September, Nvidia unveiled a new computing platform, DRIVE Thor, designed to centralise self-driving and assisted driving, along with other digital functions such as in-car entertainment.
Succeeding the successful DRIVE Orin, Nvidia founder and CEO, Jensen Huang, describes Thor as “a superchip of epic proportions… an incredible leap in deep neural network accuracy”.
The system-on-a-chip (SoC) is built on the latest central processing unit (CPU) and graphics processing unit (GPU) advances, delivering 2,000 teraflops of performance.
Self-driving superchip
“DRIVE Thor unifies traditionally distributed functions in vehicles — including digital cluster, infotainment, parking and assisted driving — for greater efficiency in development and faster software iteration,” said Huang.
“Manufacturers can configure the DRIVE Thor superchip in multiple ways. They can dedicate all of the platform’s 2,000 teraflops to the autonomous driving pipeline, or use a portion for in-cabin AI and infotainment.”
Nvidia says the SoC is capable of multi-domain computing, enabling a car to run Linux, QNX and Android simultaneously on one computer. Realistically, it could be fitted into carmakers’ 2025 models.
Danny Shapiro, head of Nvidia’s automotive business, told Reuters: “You can imagine a tremendous savings in terms of cost, in terms of reduced cabling, in terms of reduced weight, in terms of reduced energy consumption overall.”
As we noted in our “Connected car data surge” feature last year, there are pressing questions around data privacy, but there’s no doubt this technology is phenomenal.
Computational horsepower
Check out the “computational horsepower” of DRIVE Thor compared to Orin, which was itself highly rated:
For more on DRIVE Thor see this Nvidia blog.