A DATA PROCESSING SYSTEM AND METHOD FOR PERFORMING AN ARITHMETIC OPERATION ON A PLURALITY OF SIGNED DATA VALUES

摘要

The present invention provides a data processing system for performing arithmetic operations on a plurality of " p " code " n-bit " data values, and a method of operating such a data processing system. In the technique of the present invention, each code data value is aligned with each n-bit field, each code data value is sign-extended up to the most significant bit of the composite value, the sorted and sign extended data values are added, And encoding the plurality of code n-bit data values as a composite value comprising a p n -bit field. Then, the arithmetic operation is applied to the composite value to calculate an encoded result including the p n -bit field, to decode the encoded result, to extract data from each n-bit field of the encoded result Applying a decoding operation that corrects for any result caused by the addition of adjacent sign extended data values during the encoding step to produce a p final result, each of said final results comprising an arithmetic operation . According to the present invention, two or more code data values are encoded as composite values, and the composite values are processed using standard instructions as if they were SIMD instructions. This is made possible by the specific encoding of the sign data values solving the problem of overflow and underflow between two adjacent data values. Because this encoding makes standard commands available, many of the benefits of using SIMD extensions can be achieved without the need for separate hardware costs or new instructions requiring SIMD technology.