Holtek Microelectronics Inc.

George Chang

Paul Chen

Jason Chen

A high speed division method and apparatus that does not require the use of dividers are provided. The method and apparatus utilize a binary shift-add technique to simplify a divisional computation process and to enable computational error to be contained in a predetermined range, and to obtain a higher computing speed. The apparatus includes a divisor left-shift apparatus for inputting a divisor and to left-shift the divisor, a dividend left-shift apparatus for inputting a dividend and to left-shift the dividend, a left-shift controller for controlling the left-shift motion of the divisor, a quotient right-shift apparatus for inputting the indexes from a division table and then right-shift the quotient. The apparatus utilizes left-shift or right-shift to achieve a multiplication or division by 2.

Method and apparatus for high speed division

A method of compressing an integer multiplication table including the steps of first eliminating one of the two symmetrical and identical sections in the table, eliminating the products of 0 multiplier and 0 multiplicand, moving the product of the multiplier having an index of n into the location of index (n-1) for the multiplier, using one-half of the number of the largest multiplier as the largest multiplier index for the compressed multiplication table, moving into the non-continuous memory space of the compressed multiplication table by a page-filling method the product of the largest multiplier index value that is larger than the compressed multiplication table such that the multiplication table after compression can be stored on the same page.

Method and apparatus for compression of integer multiplication table

A binary multiplication method utilizing a combined table lookup and long multiplication to simplify the multiplication procedure, to improve the computational speed, and to save half of the memory space normally required. The method is executed by first moving the least significant bit (LSB) of the multiplier to another memory device before the start of the computation and then using the shortened multiplier in the multiplication operation since the multiplier is reduced by one bit, the memory space required for the multiplication table is reduced by half. The method does not require the use of a multiplying device and only needs small memory space. The manufacturing cost of a microprocessor can be reduced accordingly.

Method and apparatus for integer multiplication

A method and apparatus for determining the degree of membership in a fuzzy inference by using a subtractor and a divider without the need for a complicated multiplication, division circuit or software that is normally required in a conventional method. A fuzzy inference database for the degree of membership is first established in a microprocessor by a fuzzy inference method, the membership function in the fuzzy database has a range between 0 to 1 at full scale for the degree of membership function. When the microprocessor detects an input data, a slope distance ratio and the corresponding coordinates are determined and compared using the input data and the fuzzy database established in the microprocessor. The ratio determined by the greatly simplified method is the degree of membership function.

Growing community of inventors

Taichung, Taiwan

George Chang