FPGA - Field Programmable Gate Array

At the beginning of 1980, there were programmable logic devices, which had fast design, highly configurable and reprogrammable, but they were support only small functions. They couldn’t support large and complex function. At other end, there were spectrum of ASIC design, which support large and complex designs, but they are expensive and very time consuming. One end there was low NRE cost and second end high NRE cost. There were huge gap between digital programming devices.
To get benefit of these two programming devices, Xilinx introduced “Field Programmable Gate Array” short name “FPGA”. An FPGA have bunch of programmable logic blocks in an array with programmable switches. FPGAs are approximately 10 times less dense \ and 3 times slower than MPGAs. FPGA has two levels of programmability, each logic block can be programmed individually to perform simple logic functions and then, switches can be programmed to implement desire logic function. The key element in programmable logics are 3-input Look Up Table (LUT), multiplexer and flip-flop. The 3-input LUT is similar to PAL, used to implement combinational or Boolean equations. Each FPGA contains large number of programmable logic blocks “island” surrounded by “sea” of programmable interconnects. An FPGA have local and global programmable switches to interconnect logic blocks as well as global signals without disturbing local interconnect path.
Each FPGA vendor have different architecture, A Figure below shows general architecture of an FPGA.

Image courtesy: Xilinx