Instruction, Circuits, and Logic for Piecewise Linear Approximation
AssigneeIntel Corporation
InventorsDaniel David Ben-Dayan Rubin, Yonatan Glesner
Life of this patent
Estimating the term…
Abstract
A processor includes a linear approximator and a front end including circuitry to assign linear approximation of a nonlinear function to a linear approximator. The linear approximator includes circuitry to divide a range of values for the linear approximation into a defined number of segments, perform linear approximation for each segment, move borders between the segments to reduce discontinuity moving along segments of variable length, repeat linear approximation for each segment until convergence, and return values for the linear approximation.
Claims20
A processor, comprising:
- a linear approximator;
- and
- a front end including circuitry to assign linear approximation of a nonlinear function to the a linear approximator;
- wherein the linear approximator includes circuitry to:
- divide a range of values for the linear approximation into a defined number of equal segments;
- perform linear approximation for each segment;
- move borders between the segments such that the segments are not of equal length;
- repeat linear approximation for each segment;
- and
- return values for the linear approximation.
An apparatus, comprising a linear approximator, including circuitry to:
- divide a range of values for linear approximation of a nonlinear function into a defined number of equal segments;
- perform linear approximation for each segment;
- move borders between the segments such that the segments are not of equal length;
- repeat linear approximation for each segment;
- and
- return values for the linear approximation.
A method for linear approximation in hardware, comprising:
- divide a range of values for linear approximation of a nonlinear function into a defined number of equal segments;
- perform linear approximation for each segment;
- move borders between the segments such that the segments are not of equal length;
- repeat linear approximation for each segment;
- and
- return values for the linear approximation.
Dependencies are inferred from the claim text.