How to handle multiplication in arithmetic units

摘要

1. TECHNICAL FIELD OF THE INVENTION;How to handle multiplication in arithmetic units;2. Technical problem to be solved by the invention;Conventional multiplication methods require large registers to store intermediate results, and the lower end of the calculation results are used only for operations for rounding, which wastes a large portion of the register and increases the area of the device. To solve the problem.;3. Summary of the Solution of the Invention;Multiplying the lower bits of the multiplier and the multiplier and calculating the first rounding bit; multiplying the multiply result of the first step and the upper bits of the multiplicand and the multiplier; multiplying and calculating the second rounding bit using the multiplication result The multiplication operation is performed by dividing the final operation result and the final rounding bit into three steps from the operation result of the two steps, but using a pipeline technique for reducing the area and processing time of the device, The rounding process to reduce the error is based on the extension precision, but the rounding process can be simplified by controlling the position of the final rounding bit according to the value of the precision control signal determined at the beginning of the operation.