ARM improves ML capabilities with Ethos-U65 core
The Ethos-U55, launched in February 2020, is designed as a neural network accelerator for microcontrollers and is configurable with between 32 and 256 8×8 multiply-accumulate units (see ARM boosts machine learning for MCUs).
The Ethos-U65 offers the same performance and power efficiency as the Ethos-U55 but extends applicability to Arm Cortex-A and Arm Neoverse based systems. Although Ethos-U65 can be used alongside any Cortex-A or Neoverse CPU, it is particularly well suited to Arm CPUs with advanced vector capabilities, such as the Cortex-A55, ARM said.
The Ethos-U65 provides two different configurations of 256 and 512 MACs/cycle. It includes a dual AXI which delivers better bandwidth for weight bound networks. For more complex systems, this results in an average increase in network performance (infs/sec) of 150 percent over Ethos-U55.
The Ethos-U65 is designed for use with DRAM based systems, which leads to higher bandwidth availability. This allows Ethos-U65 to be used with more classes of embedded systems.
While using Ethos-U65 in a Cortex-A or Neoverse based system, the software flow remains the same as Ethos-U55, using TFLmicro runtime. The TFLmicro stack runs on a companion Cortex-M processor next to the NPU and handles any offload operators that cannot be executed on the microNPU itself.
With the Ethos-U65. it is possible to accelerate much larger networks, enabling applications such as real-time object detection, classification, and recognition. The Ethos-U65 uses the same software and tools as Ethos-U55.
NXP Semiconductors NV will integrate the Ethos-U65 microNPU into its next generation of i.MX applications processors.
Related links and articles:
News articles:
ARM boosts machine learning for MCUs
ARM adds Ethos cores to machine learning catalogue
NXP launches ethics initiative around AI
Aspinity raises funds for analog ML roadmap, expansion
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
