The next generation of autonomous vehicle computing improves performance and efficiency at the speed of light.
During today’s GTC keynote, NVIDIA Founder and CEO Jensen Huang unveiled DRIVE Thor, a superchip of epic proportions. The automotive-grade system-on-chip (SoC) is built on the latest CPU and GPU advancements to deliver 2,000 teraflops of performance while lowering overall system costs.
DRIVE Thor succeeds NVIDIA DRIVE Orin in the company’s product line, integrating the latest computing technology to accelerate the industrial deployment of smart vehicle technology, targeting automakers’ 2025 models.
DRIVE Thor unifies traditionally distributed functions in vehicles – including digital cluster, infotainment, parking and assisted driving – for greater development efficiency and faster software iteration.
Manufacturers can configure the DRIVE Thor superchip in several ways. They can dedicate all 2,000 teraflops of the platform to the autonomous driving pipeline, or use some for AI and in-cab infotainment and some for driver assistance.
Like the current generation of NVIDIA DRIVE Orin, DRIVE Thor uses the productivity of the NVIDIA DRIVE SDK, is designed to be ASIL D functional, and is built on a scalable architecture, so developers can seamlessly transfer their earlier software developments to the latest platform.
Swift like lightning
In addition to raw performance, DRIVE Thor offers an incredible leap in deep neural network accuracy.
DRIVE Thor marks the first inclusion of a transformer motor in the family of AV rigs. The Transformer Engine is a new component of the NVIDIA GPU Tensor Core. Transformer networks treat video data as a single perceptual image, allowing the computing platform to process more data over time.
With 8-bit floating point (FP8) precision, the SoC introduces a new automotive data type. Traditionally, AV developers experience a loss of precision when moving from 32-bit floating-point data formats to 8-bit integer data formats. FP8 precision eases this transition, allowing developers to transfer data types without sacrificing precision.
Additionally, DRIVE Thor uses updated ARM Poseidon AE cores, making it one of the highest performing processors in the industry.
DRIVE Thor is as efficient as it is powerful.
The SoC is capable of multi-domain computing, meaning it can partition tasks for autonomous driving and in-car infotainment. This multi-computing domain isolation allows concurrent time-critical processes to run uninterrupted. On a computer, the vehicle can run Linux, QNX and Android simultaneously.
Typically, these types of functions are controlled by dozens of electronic control units distributed throughout a vehicle. Rather than relying on these distributed ECUs, manufacturers can now consolidate vehicle functions using DRIVE Thor’s ability to isolate specific tasks.
All of the vehicle’s displays, sensors, and other vehicles can connect to this single SoC, simplifying what has been an incredibly complex supply chain for automakers.
Two is always better than one
If one DRIVE Thor sounds amazing, try two.
Customers can use one DRIVE Thor SoC or connect two via the latest NVLink-C2C chip interconnect technology to serve as a monolithic platform running a single operating system.
This capability gives automakers the computational headroom and flexibility to build software-defined vehicles that can be continuously upgraded through secure over-the-air updates.
Designed with the best of NVIDIA GPU technology, DRIVE Thor is truly an AV SoC of heroic proportions.