PicoCray: 9 Raspberry Pi pico RP2040 for parallel computing

The PicoCray project connects multiple Raspberry Pi Pico microcontroller modules into a parallel architecture, utilizing the I2C bus to communicate between nodes. Connect GPIO 10, 11, 12, 13 and GND to all picos in parallel. Each pico needs to be powered individually via the USB bus or the 3V3 pin. To keep the I2C bus as short as possible, you will need a pair of 4k7 ohm resistors to +3v3 on GPIO 10 and 11, ideally in the last processor in the chain. The same PicoCray code runs on all nodes, but the ground pin on one of the Pico modules indicates that it will run as a controller node. All remaining nodes run as processor nodes. Each processor node implements a random backoff technique to request addresses from the controller on the shared bus. After waiting a random amount of time, the processor will check to see if the bus is in use. If the bus is in use, the processor will return to the wait state. If the bus is not in use, the processor can request an address from the controller. Once the processor node has an address, it can send tasks from the controller node. In the sample application, these tasks involve computing the elements of the Mandelbrot set. The specific elements to be computed in a given task are assigned by the controller node, which then collects the results from each processor node and aggregates the results for display. The name of the project was inspired by Seymore Cray. Our biography of the father of supercomputers tells his story, including why the Cray-1 supercomputer is called "the world's most expensive loveseat."