NVIDIA Unveils Smaller, More Efficient Jetson Modules
Autonomous robots are rapidly moving beyond research labs and into warehouses, factories, stores, and other real-world environments. But running the sophisticated AI models needed to perceive surroundings, understand instructions, and make decisions requires a lot of computing power. NVIDIA is attempting to address that challenge with two new Jetson modules that bring its Blackwell-powered Thor architecture to smaller, more power-efficient edge AI systems.
The new Jetson T3000 sits near the high end of the lineup, delivering 865 FP4 teraflops of AI compute in a module that is roughly half the size and consumes half the power of the T5000. It combines a Blackwell GPU with an eight-core Neoverse Arm CPU, 32GB of LPDDR5X memory, 273GB/s of memory bandwidth, and 25 GbE connectivity.
T3000 performance benchmarks (📷: NVIDIA)
Despite the reduction in size and power requirements, the T3000 can achieve inference performance similar to the T5000 for multimodal workloads. These include...
Copyright of this story solely belongs to hackster.io. To see the full text click HERE
