论文部分内容阅读
根据实际科研工作的需要 ,进行浮点部件 NRS FPU中基本算术运算算法的研究与实现。对完成指数和尾数加减的回绕进位加法器算法进行了详细的分析。同时 ,对尾数乘运算的改进BOOTH算法进行改造 ,改进了 BOOTH的译码表 ,在 BOOTH译码过程中省掉了数据符号位的保存和参与 ,简化了乘运算实现的逻辑复杂度。最后 ,提出了新的扩展 GOLDSCHMIDT算法 ,有效减少 GOLDSCHMIDT算法的乘运算次数 ,提高了运算速度。另外 ,对浮点的加、减、乘、除等基本运算设计了实现流程
According to the actual needs of scientific research, research and implementation of the basic arithmetic operation algorithm in floating point NRSFPU. The wraparound adder algorithm that completes addition and subtraction of exponent and mantissa is analyzed in detail. At the same time, the modified BOOTH algorithm of mantissa multiplication is modified to improve the decoding table of BOOTH. The saving and participation of data symbol bits in BOOTH decoding process are simplified, which simplifies the logic complexity of multiplication operation. Finally, a new extended GOLDSCHMIDT algorithm is proposed, which can effectively reduce the multiplication of GOLDSCHMIDT algorithm and improve the computing speed. In addition, the floating point addition, subtraction, multiplication and division and other basic operations designed to achieve the process