There are currently several professional and maker solutions for Raspberry Pi boards that are designed for cluster computing — including MiniNodes’ 5 Node Raspberry Pi 3 CoM Carrier Board, PicoCluster’s scalable Pico series nodes, and Pimoroni’s Cluster HAT, to name just a few. We can now add the Turing Pi Clusterboard to that list, which is a compact solution that leverages the Raspberry Pi Compute Module.
The Turing Pi Clusterboard is a Mini ITX-sized board that can take advantage of seven Compute Modules and is geared for applications that support the Kubernetes software ecosystem, Docker, Jupyter Notebook, machine learning TensorFlow and Caffe software, as well as serverless stacks.
As for its specs, thhe 170 x 170mm Turing Pi features seven sockets that support the Raspberry Pi Compute Module 3/3+, seven microSD card slots, an HDMI port, MIPI DSI connector, and a 3.5mm audio jack. It also packs two MIPI CSI connectors (camera/IF), Gigabit Ethernet port and onboard switch, 8X USB ports, and can either use a 12V-20V DC jack or ATX power supply.
There’s very little information on the Turing Pi product page at this point, with no tutorials or software installation, only that it can be pre-ordered for $128 and will ship sometime this fall. You will have to supply your own Raspberry Pi Compute Modules, microSD cards, and power supply, which can run anywhere from $25 to $35 for each board alone, sending that price tag well over $300.