Implementation of a Speed Optimized Pipelined 32-Bit Floating Point Multiplier for Performance – Critical VLSI Systems