Speaker
Description
As particle detectors advance to reach higher and higher data rates, the demand on frontend and readout electronics grows. Commercial FPGA development kits are designed to demonstrate device capabilities but typically do not expose the raw I/O needed to build custom data acquisition systems, leaving many experiments to choose between expensive custom designs and inadequate off-the-shelf hardware. We present KrIO, a low-cost, high-throughput FPGA platform built around the AMD Kria K26 System-on-Module. KrIO provides a standardized base for semi-custom DAQ systems that can be integrated with an experiment without advanced board-level engineering. High-speed readout is delivered through two Gigabit Ethernet ports and a four-lane PCIe interface, while firmware development is accelerated by AMD’s mature toolchain as well as HLS4ML to deploy real-time ML inference in the data path. KrIO makes it possible to build a high-performance data acquisition system with integrated real-time ML/AI on a limited timescale and budget, for experiments and detector R&D projects of all sizes.