I spotted that the open hardware ~$100 Beaglebone AI-64
Now supports Dockerised Ubuntu 22.04 under Yocto
Does anyone have ROS running on it yet? It seems like it might be good for robotics?
Memory
4GB double-data-rate 4 (DDR4) RAM
16GB embedded MultiMediaCard (eMMC) onboard flash storage
Processor
Dual Arm® Cortex®-A72 microprocessor subsystem at 2GHz C7x plus MMA (8 TOPS) and two C66x floating-point very long instruction word (VLIW) digital signal processors (DSPs)
Three dual Arm Cortex-R5 coprocessors
Two six-core programmable real-time unit and industrial communication subsystem (PRU-ICSSG)
PowerVR® Series8 GE8430 3D graphics processing unit (GPU)
Accelerated video codecs (two 1080p30 H.264 encode, eight 1080p30 H.264/H.265 decode)
- One dual-core 64-bit Arm Cortex-A72 microprocessor subsystem at up to 2.0 GHz and up to 24K DMIPS (Dhrystone Million Instructions per Second)
- Up to three Microcontroller Units (MCU), based on dual-core Arm Cortex-R5F processor running at up to 1.0 GHz, up to 12K DMIPS
- Up to two TMS320C66x DSP CorePac modules running at up to 1.35 GHz, up to 40 GFLOPS
- One C71x floating point, vector DSP running at up to 1.0 GHz, up to 80 GFLOPS
- One deep-learning MMA, up to 8 TOPS (8b) at 1.0 GHz
- Up to two gigabit dual-core Programmable Real-Time Unit and Industrial Communication Subsystems (PRU_ICSSG)
- Two Navigator Subsystems (NAVSS) for data movement and control
- One multi-pipeline Display Subsystem (DSS) with one MIPI® Display Serial Interface Controller (DSI) and shared MIPI D-PHY Transmitter (DPHY_TX), one Embedded DisplayPort Transmitter (EDP) with shared Serializer/Deserializer (SERDES), and two MIPI Display Pixel Interface (DPI) ports
- Two Camera Streaming Interface Receivers (CSI_RX_IF) with dedicated MIPI D-PHYs (DPHY_RX)
- One Camera Streaming Interface Transmitter (CSI_TX_IF) with MIPI D-PHY Transmitter (DPHY_TX) shared with DSI
- One Vision Processing Accelerator (VPAC) with image signal processor
- One Depth and Motion Processing Accelerator (DMPAC)