Method and apparatus for improving the precision or area of a memory table used in floating-point computations
US5768171A · kind A · utility
Assignee
Inventor
Key dates
| Filing date | Jun 28, 1996 |
| Grant date | Jun 16, 1998 |
| Priority date | — |
| Expiry date | Jun 28, 2016 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F2101/12
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A method and apparatus for calculating a value of a function f(x) for a given operand x. A memory, such as a Read Only Memory, is used to stored precalculated values for some bits of the function f(x) for a plurality of values of the operand x. Bits of the operand x are used to generate bits for the value of the function f(x) in two ways: 1. to address the ROM to generate some bits for the value of f(x), and 2. as inputs to combinational logic to generate an additional bit of the value of f(x). Because one bit of the value for f(x) is generated external to the ROM, the size of the ROM can be reduced without sacrificing accuracy. Alternatively, the ROM can be used to store an additional bit of precision for values of f(x), thereby increasing the accuracy of these values.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.