FPGAs get real

1 min read

What fpgas need from a real time operating system. By Colin Walls.

Until a couple of years ago, the idea of mentioning a real time operating system (RTOS) and an fpga in the same sentence would not have occurred to anyone. The average software engineer had little idea what an fpga was and fpga designers could not have cared less about RTOSes. So, what changed? Two things did: * The major fpga vendors announced new devices which incorporated powerful processor cores (ARM or PowerPC), and * A number of soft core processors (devices designed to be instantiated into the fpga fabric) were announced (notably Nios from Altera and MicroBlaze from Xilinx). Until then, embedded software engineers had either worked with discrete processors on boards, with memory and peripheral devices, or they had been involved in system on chip [SoC] designs. Now there was a third context in which they might develop code: often termed the ‘field programmable system on chip’, or FPSoC.