Hackster is hosting Hackster Holidays, Ep. 7: Livestream & Giveaway Drawing. Watch previous episodes or stream live on Friday!Stream Hackster Holidays, Ep. 7 on Friday!

OV Tech's Pi.MX8 Offers a Raspberry Pi Compute Module 4 Alternative with On-Board AI Acceleration

Built around NXP's i.MX 8M Plus system-on-chip, this open-hardware CM4-compatible system-on-module includes a 2.3 TOPS NPU.

Gareth Halfacree
12 months ago β€’ Machine Learning & AI / HW101

German open hardware specialist OV Tech is preparing to launch a drop-in replacement for those building around the Raspberry Pi Compute Module 4 (CM4) system-on-module: the Pi.MX8, powered by the NXP i.MX 8M Plus system-on-chip and its integrated accelerator for on-device machine learning and artificial intelligence (ML and AI.)

"Pi.MX8 is an open source compute module based on the popular, dependable i.MX 8M Plus SoC series from NXP," OV Tech's Lukas Henkel explains. "The board is compatible with the popular Raspberry Pi CM4 module standard both in form factor and electrical specifications. NXP's guaranteed long-term processor support of 15 years combined with the expertise of a large customer and user base for the i.MX8 SoC gives you the security you need to build your system on."

Designed to slot into any carrier board designed for the Raspberry Pi Compute Module 4, the Pi.MX8 features four 64-bit Arm Cortex-A53 cores running at up to 1.8GHz and a Cortex-M7 real-time core running at up to 700MHz, a Vivante GC700UL graphics processor, a neural network coprocessor with a claimed 2.3 tera-operations per second (TOPS) compute performance, a choice of 1GB, 2GB, 4GB, or 8GB of LPDDR4 RAM, and a choice of eMMC storage up to 32GB or a microSD card β€” both of which can be used simultaneously, if desired.

Other features of the design includes gigabit Ethernet with IEEE 1588v2 and SyncE support, Raspberry Pi-compatible general-purpose input/output (GPIO) capabilities, and a PCI Express Gen. 3 interface. Optional upgrades include a Murata LBEE5PK2BC-based Wi-Fi and Bluetooth 5.2 module and a Google Coral Edge TPU machine-learning accelerator β€” which, if installed, connects over the PCIe lane. The module also includes Camera Serial Interface (CSI) connectivity for imaging modules, USB connectivity, and HDMI for video and audio output.

For those not already in the Compute Module 4 ecosystem, OV Tech is also offering "a complete i.MX8 system" which pairs the Pi.MX8 module with a carrier board breaking out Ethernet, two USB, and HDMI ports plus the microSD slot, with "one of the smallest enclosures available" for the form factor. "Inside the enclosure, an FPC [Flexible Printed Circuit] connector and FPC cable are available to route PCIe, 5V, and 3.3V power as well as CSI interfaces to your custom hardware," Henkel adds.

The Pi.MX8 and its carrier bundle are due to begin a crowdfunding campaign on Crowd Supply in the near future, at an as-yet unannounced price point; OV Tech has confirmed that the module's design files will be published to GitHub under an unspecified open source license when the funding campaign concludes.

Gareth Halfacree
Freelance journalist, technical author, hacker, tinkerer, erstwhile sysadmin. For hire: freelance@halfacree.co.uk.
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles