“Performance + Time = Memory” is a strange
formula, but when understood, it often results in
significantly lower cost implementations with Xilinx
devices. While this is vital for digital signal processing (DSP) applications, it can be applied quite effectively to many different types of designs, and is particularly useful for applications that use the range of Xilinx Spartan devices, for which cost savings are always welcome.

This paper investigates the formula and why it is so applicable to Xilinx devices. It also presents a case study to consider, as well as potential solutions.