What is an FPGA?

By Cabe Atwell Mar 8, 2022 10:00am

While the concept isn’t new, field-programmable gate arrays (FPGAs) have gained increasing notoriety, especially in embedded system development and deep neural networks for artificial intelligence applications. In 1985, Xilinx co-founders invented the XC2064, the first market-worthy, field-programmable gate array with both programmable gates and programmable interconnects between gates. The precursors to FPGAs were programmable read-only memory (PROM) and programmable logic devices (PLDs).

What’s in a name?

What makes FPGAs so desirable is their flexibility. FPGAs consist of an array of programmable logic blocks which can serve as simple logic gates (e.g., AND and OR) or be configured to perform complex combinational functions. As their name suggests, they are also field-programmable and are designed to be configured by the user. To further understand FPGAs, let’s break down their many attributes, starting with the name.

Field-Programmable: After manufacture, the user has the ability to program and reprogram the hardware circuit or circuits of an FPGA to perform one or more logical operations.
Gate + Array: An FPGA is an integrated circuit or sets of circuits or arrays. These arrays are programmable logic gates or configurable logic blocks (CLBs), memory, etc.

illustration of an earlier FPGA — An early FPGA from Xilinx (Raimond Spekking)

Other Attributes of FPGAs

Now that we’ve covered the basics of the name, here are some additional characteristics of FPGAs:

Similar to application-specific integrated circuits (ASICs), FPGA configuration is specified using a hardware description language (HDL).
Structurally, FPGAs consist of reconfigurable interconnects that enable blocks to be wired together.
FPGAs enable flexible, reconfigurable computing, as most enable the user to reprogram the hardware circuit with different logic functions. This is especially useful in computer software.
Typically, the logic blocks in FPGAs include memory components, which might be simple or complex.

FPGA Advantages Over ASIC

FPGAs have taken the main stage when it comes to embedded system development and hardware prototyping. Here, their iconic flexibility comes into play by giving developers the ability to conduct trials and test variables and make subsequent changes to configuration files before they finalize the system architecture. Here are some additional advantages to FPGA:

Functionality can be updated after manufacture
Ability to partially or fully reconfigure a portion of the design
Low, non-recurring engineering costs
Ability to prototype before going to market
Shorter time to market

FPGA Disadvantages Over ASIC

In comparison to application-specific integrated circuits, FPGAs typically possess certain inherent disadvantages, though continuous development and advancements will likely minimize these over time.

Higher cost per unit
Can be more vulnerable to tampering
Generally are larger in size
Require more power
Are slower

FPGA Uses & Industries

Today, there are numerous uses and applications for FPGA technology, with more in development every day. Here’s a taste of FPGAs in use across various industries.

Aerospace: Radar and sonar
Automotive: Autonomous driving and advanced driver-assistance systems
Broadcast: Broadcast Target Design Platforms
Consumer: Computers and smartphones
Industrial: Motor control, machine vision, video surveillance and data processing
Medical: Diagnostic, monitoring and therapy devices
Networking: Digital signal processing, neural networks, scientific instruments and medical devices
Telecommunications: Packet processing and switching and optical transport network functions

.