Browse Prior Art Database

Method and System for Estimating Power Consumption of Microprocessor at Different Operating Points

IP.com Disclosure Number: IPCOM000196813D
Publication Date: 2010-Jun-17
Document File: 4 page(s) / 41K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method and system for estimating power consumption of a microprocessor at different operating points. A power model is provided that takes power consumption values at different operating points as input to estimate power consumption at a target operating point.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 4

Method and System for Estimating Power Consumption of Microprocessor at Different

Operating Points

A method and system is disclosed for estimating power consumption of a microprocessor at different operating points. A power model is provided that takes power consumption values at different operating points as input to estimate power consumption at a target operating point.

In accordance with the method and system disclosed herein, the power model is calibrated using calibration vectors and known power consumption values at known operating points as shown in fig.1. Subsequently, the power model is employed to estimate power consumption at a target operating point.

Figure 1

The power model is calibrated by measuring at least two power consumption values using at least two operating points. In an embodiment, the power consumption values may be selected so as to cover entire span of the power consumption of the microprocessor. For example, the power consumption value at an idle situation and the power consumption value at maximum power workload.

Thereafter, regression analysis is used to derive the power model. For example, linear regression is used to derive an order N-1 polynomial equation to fit N points. Further, each workload may use a different regression method, if desired.

1

[This page contains 1 picture or other non-text object]

Page 2 of 4

Fig. 2 illustrates a 2nd order polynomial power model.

Figure 2

Using the power model as illustrated in fig. 2, high power and low power workload are measured at f1, f2 and f3. Further, coefficients a, b, c, q, r and s are determined using linear regression.

Once the power model is derived, current power at current...