Method and apparatus for performing numerical calculations

摘要

There is provided a method of processing an iterative computation on a computing device comprising at least one processor. Embodiments of the method comprises performing, on a processor, an iterative calculation on data in a fixed point numerical format having a scaling factor, wherein the scaling factor is selectively variable for different steps of said calculation in order to prevent overflow and to minimize underflow. By providing such a method, the reliability, precision and flexibility of floating point operations can be achieved whilst using fixed point processing logic. The errors which fixed-point units are usually prone to generate if the range limits are exceeded can be mitigated, whilst still providing the advantage of a significantly reduced logic area to perform the calculations in fixed point.

主权项

1. A method of processing an iterative computation on a computing device comprising at least one processor by performing, on a stream processor, an iterative calculation on data in a fixed point numerical format having a scaling factor that is selectively variable for different steps of the calculation in order to prevent overflow and to reduce underflow, the method comprising:
inputting a plurality of arrays of data to the stream processor; scaling each array to a common scaling factor selected to prevent overflow and reduce underflow; calculating, by the stream processor, a step of a numerical computation in a fixed point format using the scaled array data; applying a predetermined change to the scaling factor with each step of the iterative calculation, the value of the predetermined change being the same for each step of the iterative calculation; updating, using a processor, the scaling factor based upon the resulting values of the calculation step; and repeating the steps of inputting, scaling, applying, and updating for the next step of the numerical computation using the updated scaling factor.