Turing Pi

Plug & Play Raspberry Pi cluster.
Host Kubernetes, Docker, Serverless locally.

$189

Out of stock

Notify Me
turingpi-front
turingpi-back

Specs

Specs:
  • 1 Gbps Ethernet
  • Up to 28 cores
  • 7 x 40 pin GPIO
  • HDMI, Audio jack, ATX
  • 8 x usb
Features:
  • Flash Mode
  • Boot mode
  • Nodes power management
  • Real-time clock (RTC)
  • Cluster management bus (I2C)

Software

Turing Pi supports Kubernetes software ecosystem, as well as machine learning and serverless stacks
Kubernetes
Tensor Flow
Docker
PyTorch
Jupyter Notebook
OpenFaaS
Caffe
MXNet

FAQ

What can I do with Turing Pi?
  • Home server (homelab) and cloud apps hosting
  • Learn Kubernetes, Docker Swarm, Serverless, Microservices on bare metal
  • Cloud-native apps testing environment
  • Learn concepts of distributed Machine Learning apps
  • Prototype and learn cluster applications, parallel computing, and distributed computing concepts
  • Host K8S, K3S, Minecraft, Plex, Owncloud, Nextcloud, Seafile, Minio, Tensorflow
Does Turing Pi includes compute modules?

No, the price is only for the Turing Pi board

How to the compute modules communicate with each other?

The nodes interconnected with the onboard 1 Gbps switch. However, each node is limited with 100 Mbps USB speed. Also, there is an I2C bus to exchange some technical information between nodes, including Real-Time Clock (RTC).

From where Turing Pi boots OS?

You can boot the OS either from eMMC, SD card or netboot.

Can I flash compute modules through the board?

Yes, you can flash a compute module using a top/master node.

How do the NIC, Ethernet, USB, HDMI, and audio ports work?

There are 8 USB on the board. Each pair of USB connected to a particular node. 2x USB routed to the top/master node, 2x to the second node, 2x to the fourth node, 2x to the 6th node. HDMI and audio connected with a top/master node.

NIC - There is an 8 port switch on the board. Each port goes to each node plus one uplink.

Do all the slots need to be filled in?

Turing Pi works with any amount of nodes. You can start with a couple of nodes and scale when needed.