Fast divider for a constant divisor

Claim 1

A divider for dividing a dividend by a divisor to calculate a quotient and a remainder, said divisor being a natural number, each of said dividend, said quotient, and said remainder being an integer which is not less than zero, said dividend being L bits long, where L represents a first predetermined natural number which is not less than two, said divisor being a constant which is M bits long, where M represents a second predetermined natural number which is not less than two, comprising: first through N-th comparing means supplied with said dividend in common and with first through N-th predetermined constants, respectively, for comparing said dividend with said first through said N-th predetermined constants, where N represents a third predetermined natural number which is not less than two, said third predetermined natural number N being related to said first and said second predetermined natural numbers L and M, respectively, by N=2 LM -1, an n-th predetermined constant being equal to n times said divisor, where n is variable between 1 and N, both inclusive, said first through said N-th comparing means producing first through N-th comparison result signals indicative of first through N-th comparison results, respectively; decoding means connected to said first through said N-th comparing means for decoding a combination of said first through said N-th comparison result signals into a decoded signal which is divided into first and second partial decoded signals, said first partial decoded signal being equal to said quotient, said second partial decoded signal being equal to lower M bits of a product of said quotient and said divisor, the lower M bits of said product being arranged from a least significant bit 2 0 to an M-th bit 2 M-1 as counted from said least significant bit 2 0 ; and operation means supplied with said dividend and connected to said decoding means for carrying out an operation on said second partial decoded signal and lower M bits of said dividend that are arranged from a least significant bit 2 0 to an M-th bit 2 M-1 as counted from said least significant bit 2 0 , said operation means thereby producing said remainder.