VLSI floating-point processors

The advance of VLSI technology has been the enabling factor in the appearance of VLSI circuits handling floating-point arithmetics. These circuits have found their way into many number-crunching applications such as telecommunications, seismic energy exploration, radar, medical imaging, graphics and simulation. Because of the different requirements for different applications, some processors have a rich repertoire of functions but rather low performance, while some processors aim at having the highest throughput for the most frequent operations such as multiply and add. This paper will review the architecture, the technology, and the design techniques for the current VLSI floating-point processors; and it will also report a high performance chipset (^∗) implementing complete basic arithmetic functions.