We propose a floating–point representation to deal efficiently with arithmetic operations in codes with a balanced number of additions and multiplications for FPGA devices. The variable shift operation is very slow in ...