NVIDIA's next-generation GPU architecture Blackwell unveiled to help automate AI tasks

NVIDIA recently released its latest GPU architecture, Blackwell, which has significantly improved energy efficiency and AI computing performance compared to the previous generation Ampere architecture. The Blackwell architecture supports higher INT8/FP8 computing power, making it more suitable for deep learning inference tasks in autonomous driving. Specifically, the Orin-X chip is equipped with the Ampere architecture, with an FP32 computing power of 5.2 TFLOPS, and is suitable for medium-complexity AI tasks; the Thor-X chip is equipped with the Blackwell architecture, with an FP32 computing power of 9.2 TFLOPS, and is suitable for For multi-camera scenarios; the Thor-X-Super chip is also equipped with Blackwell architecture and has 18.4 TFLOPS of FP32 computing power. It is suitable for fully autonomous driving systems and can handle highly complex AI tasks.