Paper Title
Multi-Mode Floating Point Multiplier With Parallel Operations
Abstract
Most modern processors have hardware support for single precision and double precision floating point
multiplication. For many scientific computations like climate modeling, computational physics and computational geometry
this support is inadequate. They impose the use of quadruple precision arithmetic because it provides twice the precision of
double precision format. The proposed design performs four single precision multiplications in parallel, or two double
precision multiplications in parallel, or one quadruple precision multiplication. The throughput is increased by a factor of four
for single precision multiplication and by two for double precision multiplication