History Software Engineering

Field-Programmable Gate Array (FPGA) – 1985 AD

Return to Timeline of the History of Computers


Field-Programmable Gate Array

Ross Freeman (1948–1989), Bernard Vonderschmitt (1923–2004)

Many kinds of calculations can be implemented in either hardware or software. Typically, hardware is faster but more complicated, while software is slower but easier to create and debug. That’s because hardware typically has many circuits and wires that perform the calculation in parallel. Software, in contrast, runs as a series of instructions within a computer’s CPU: the same circuits and wires are repurposed over the course of the calculation for many different uses. Software remains more popular than hardware for solving problems, because it’s typically easier to develop and change.

But what if hardware could be programmed just like software? Programmable hardware could implement special algorithms for video processing. It could run AI algorithms for image recognition at high speed, and presumably with less power. Programmable hardware could also be used to replace complicated circuit board designs with hundreds of individual components with a single, programmable chip.

That’s the idea of a field-programmable gate array (FPGA). The chips contain programmable logic cells that can be connected as needed. Once they are wired up, the gates can work like the circuits inside any other silicon chip—with one big difference. If the wiring diagram isn’t correct, or if it needs to be changed, the program can be erased and the FPGA reprogrammed with a new configuration.

FPGAs are typically more expensive to purchase and program than application-specific integrated circuits (ASICs). But because they can be programmed in the lab and then reprogrammed as necessary, they make innovation dramatically cheaper and faster—especially in applications where only a few integrated circuits are needed, such as a prototype. Otherwise, the cost of replacing a buggy circuit would be prohibitive, like on a spacecraft. That’s why NASA used FPGAs on its Mars Curiosity rover.

Ross Freeman and Bernard Vonderschmitt cofounded Xilinx® in 1985 and created the first commercially viable FPGAs, winning Freeman a place in the National Inventors Hall of Fame.

SEE ALSO Silicon Transistor (1947), First Microprocessor (1971), Verilog (1984)

The FPGA at the center of this circuit board can be programmed to pulse lights, spin motors, and synthesize music.

Fair Use Source: B07C2NQSPV