Embedded AI Hardware

Hardware
, ,
1

This topic is to note various embedded AI hardware devices.

Comparison of Linux Based AI Devices

Device Tops CPU GPU/NPU
Jetson AGX Orin 200-275
Jetson Orin NX 70-100
Jetson Orin Nano 20-40
Particle Tachyon 12 Qualcomm Kyro CPU Octa-core
1 TI AM62A single to quad-core Cortex-A53 C7x256v + MMA
BeagleY-AI 4 TI AM67A Quad-core
8 TI AM68A C7x + MMA
32 TI AM69A 4X C7X DSP + 4X MMA
Variscite VAR-SOM-MX93 0.5 NXP i.MX93 Dual-core Cortex-A55 NXP eIQ® Neutron NPU
2 NXP i.MX95 NXP eIQ® Neutron NPU
2.3 NXP i.MX 8M Plus
2

The newly developed technologies are as follows: (1) A dynamically reconfigurable processor (DRP)-based AI accelerator that efficiently processes lightweight AI models and (2) Heterogeneous architecture technology that enables real-time processing by cooperatively operating processor IPs, such as the CPU. Renesas produced a prototype of an embedded AI-MPU with these technologies and confirmed its high-speed and low-power-consumption operation. It achieved up to 16 times faster processing (130 TOPS) than before the introduction of these new technologies, and world-class power efficiency (up to 23.9 TOPS/W at 0.8 V supply).

3

The board is equipped with the DRP-AI3 NPU, which delivers AI inference performance of up to 80 TOPS on an INT8 basis useful for power-sensitive AI image processing tasks. The hardware also supports acceleration for OpenCV, enhancing its capabilities for detailed image processing.

4