Skip to content

pico2-ice

The pico2-ice is a development board from TinyVision.ai that combines a Raspberry Pi RP2350B microcontroller with a Lattice iCE40UP5K FPGA on a single compact PCB. It is the next-generation hardware platform for the Bitstream Evolution project, enabling on-board microcontroller-driven evolution and parallel evaluation across multiple boards via the iCEFARM system.

Official Documentation Purchase Hardware Setup Guide

Specifications

Feature Details
FPGA iCE40UP5K — 5,280 LUTs, 1 Mbit SPRAM, 120 Kbit DPRAM, 8 multiplier blocks
Microcontroller RP2350B — dual Arm Cortex-M33 + dual RISC-V Hazard3 cores
External Memory 8 MB low-power qSPI SRAM
Flash 4 MB SPI Flash (independent storage for FPGA and RP2350B)
GPIO All RP2350 + 32 FPGA GPIO on 0.1" headers (Pmod-compatible)
LEDs 2 RGB LEDs (one per processor)
Buttons 2 pushbuttons
PCB 4-layer with dedicated ground plane

Advantages Over the iCEstick

  • 4x more logic --- the UP5K provides 5,280 LUTs compared to the HX1K's 1,280 logic cells, offering significantly more resources for evolved circuits.
  • On-board microcontroller --- the RP2350B could in principle run the evolutionary algorithm locally, eliminating the need for a separate host PC and external Arduino.
  • Programmable clock --- the RP2350 supplies the FPGA clock, allowing software-controlled frequency adjustment.
  • Direct FPGA programming --- the microcontroller can load bitstreams onto the iCE40UP5K, enabling streamlined evolutionary runs.
  • Array-friendly --- the compact form factor and on-board MCU make it suitable for deploying multiple boards in the iCEFARM parallel evaluation system.

Bootloader Mode

To flash new firmware onto the pico2-ice, you need to enter bootloader mode. Hold the SW1* button down while connecting the board to USB and — the board will appear as a USB mass storage device on your computer, ready to receive a new .uf2 firmware file.

Project Use

The pico2-ice is the hardware platform for the Bitstream Evolution (pico2-ice) project and the iCEFARM parallel evaluation system. Adopted in 2025 as part of an ARI-funded joint project between Vivum AI and Rose-Hulman Institute of Technology, it enables research into partial reconfiguration of FPGAs for autonomous control systems and larger-scale evolutionary experiments.