Chip manufacturer Nvidia has introduced a new high-performance chip for autonomous driving: Drive Thor.
Of the Nvidia Drive Thor is a new generation centralized computer chip for autonomous driving. The chip has a performance of up to 2,000 teraflops, which means 10 to the power of 12 (floating point number) operations per second. In this way, the computer in the car can calculate faster, for example process sensor data faster or calculate the navigation.
The Drive Thor should not only accelerate autonomous driving, but also allow parking, the driver’s attention check and operate the displays in the car and provide entertainment. It’s all unified in the architecture of the system, which coincides with decreasing overall costs.
The chip is the successor of the Nvidia Drive Orin from 2019, which started with 254 terflops. The Thor chip also has an AI capability that is accelerated by a factor of 9. DRIVE Thor with MIG support for graphics and calculations enables domain isolation to run, allowing simultaneous time-sensitive processes to run without interruption. The chip also supports multi-domain computing and IVI. Thor also allows the transition to 8-bit without sacrificing accuracy. The DRIVE Thor SoC and the AGX board should comply with ISO 26262.
The system-on-a-chip (SoC) concept called Thor should be available to manufacturers from 2025 and has already found a buyer: the Chinese manufacturer Geely respectively his child Zeekr will integrate the chip into its models. Its production begins 2025.
The new chip also uses the latest NVLink C2C chip interconnect technology while running multiple operating systems. NVLink-C2C’s advantage is the ability to share, schedule, and distribute work across the link with minimal overhead. This allows manufacturers the computing power to software-defined vehicles (SDV) that are updatable through secure over-the-air software updates.