By KL Lim
The NVIDIA Orin low-power system-on-chip based has set new records in AI inference in its debut on the MLPerf benchmark.
A pre-production version of Ampere architecture-based chip led in five of six performance tests on edge AI. It is up to 5x faster than the previous generation Jetson AGX Xavier and delivers an average of 2x better energy efficiency.
Servers and devices with NVIDIA GPUs including Jetson AGX Orin were the only edge accelerators to run all six MLPerf benchmarks.
NVIDIA Orin is available in the NVIDIA Jetson AGX Orin developer kit for robotics and autonomous systems. The Jetson platform for AI inference is used by more than 6,000 customers including Amazon Web Services, John Deere, Komatsu, Medtronic and Microsoft Azure.
China’s largest electric vehicle (EV) maker BYD has announced that it will use the Orin-based Drive Hyperion architecture for its next-generation automated EV fleets.
Orin is also used in NVIDIA Clara Holoscan for medical devices by system makers and researchers to develop next generation AI instruments.
The MLPerf benchmarks are backed by organisations such as Amazon, Arm, Baidu, Dell Technologies, Facebook, Google, Harvard, Intel, Lenovo, Microsoft, Stanford, and the University of Toronto.
MLPerf’s diverse tests cover today’s most popular AI workloads and scenarios, giving users confidence the benchmarks will reflect performance they can expect across the spectrum of their jobs.
“Overall, NVIDIA with its partners continued to show the highest performance and broadest ecosystem for running all machine-learning workloads and scenarios in this fifth round of the industry metric for production AI,” said Dave Salvator, Senior Manager of Accelerated Computing at NVIDIA.