JPS54162936A - Data processor - Google Patents

Abstract

PURPOSE:To obtain a high-speed and accurate quotient under a simple control for the system through which the approximate answer of the division is obtained by repetition of multiplication by obtaining the approximate answer in the extra bit number rather than the necessary significant digit number and then deciding the presence or absence of correction. CONSTITUTION:The approximate answer is obtained in n-bit more than necessary significant digit number m-bit, and such correction is given to secure the quotient Q for the higher-rank m-bit which is the result obtained by adding 1 to the m+n-bit position from the higher rank of the approximate answer. In such case, the contents of extra n-bit is checked through correction decision circuit 20 to decide the presence or absence of correction. Quotient Q is defined as the final quotient at circuit 20 in case the contents of the n-bit is not all 0 nor 1, thus operation end signal 202 being delivered. In case the contents of the n-bit is all 1 or 0, the necessity of correction is informed to corrective operation control circuit 21. Circuit 21 controls control parts 22-25 for each of the input data, division, multiplication and subtraction respectively to carry out an operation of ¦dividend¦ - Quotient x divisor¦= R. Thus, Q or the value obtained by subtracting 1 from the lowest-rank bit position of Q are defined as the final quotient in the case of R>=O and R<O respectively.