All is Not SRAM - A survey of flash, antifuse, and EE programmable logic

There are PCs, but there are also Macs. There is Windows, but there is also Linux. There are automobiles, but there are also motorcycles. For every mainstream movement, there is an alternative subculture – the twenty-percent faithful who traverse the road less traveled. In FPGA, the mainstream is SRAM. The two largest FPGA suppliers have engaged in a years-long battle for SRAM supremacy, and the result is a mature, well-developed, and highly capable technology. SRAM devices have raced to ever higher densities, faster clock speeds, more robust design environments, and more complete IP portfolios, fueled by explosive competition, particularly in the telecom market. At the same time, however, other technologies have quietly flourished on the sidelines, developing differentiators that may become increasingly important as the programmable logic market grows and diversifies.

Depending on your application, there are several compelling reasons you might want to consider alternatives to SRAM-based FPGAs. The SRAM FPGA offerings on the market right now do certain things very, very well. The latest FPGA devices from Altera and Xilinx offer a very competitive point in the price/performance/density space. If you did your measurements in some bizarre dimensions like gate-megahertz-per-dollar at lowish volumes you’d find SRAM FPGAs near the top of the heap. For many designs, however, alternatives such as flash, antifuse, EE, and the venerable CPLD can offer a better mix of features, performance, and capability. [more]