Adaptive real-time methodology for optimizing energy-efficient computing
AssigneeLos Alamos National Security, LLC
InventorsChung-Hsing Hsu, Wu-Chun Feng
Life of this patent
Estimating the term…
Abstract
Dynamic voltage and frequency scaling (DVFS) is an effective way to reduce energy and power consumption in microprocessor units. Current implementations of DVFS suffer from inaccurate modeling of power requirements and usage, and from inaccurate characterization of the relationships between the applicable variables. A system and method is proposed that adjusts CPU frequency and voltage based on run-time calculations of the workload processing time, as well as a calculation of performance sensitivity with respect to CPU frequency. The system and method are processor independent, and can be applied to either an entire system as a unit, or individually to each process running on a system.
Claims19
A computer-implemented method for optimizing energy consumption for an active computing event, comprising:
- measuring a workload processing rate for a range of central processing unit (CPU) frequencies;
- and
- repeating the following steps until the event is no longer active:
- computing a run-time performance sensitivity factor;
- wherein said run-time performance sensitivity factor is computed relative to a change in CPU frequency;
- determining a target frequency based on said run-time performance sensitivity factor;
- wherein said target frequency comprises a minimum frequency having a predicted performance slowdown that does not exceed a maximum possible performance slowdown;
- and
- setting the CPU frequency to said target frequency;
- and
- updating the workload processing rate for said target frequency;
- wherein energy consumption is optimized under a performance slowdown constraint.
A system for optimizing energy consumption for an active event, comprising:
- a computer;
- and
- programming on said computer for executing the steps of:
- measuring a workload processing rate for a range of central processing unit (CPU) frequencies;
- and
- repeating the following steps until the event is no longer active:
- computing a run-time performance sensitivity factor;
- wherein said run-time performance sensitivity factor is computed relative to a change in CPU frequency;
- determining a target frequency based on said run-time performance sensitivity factor;
- wherein said target frequency is a minimum frequency whose predicted performance slowdown does not exceed a maximum possible performance slowdown;
- and
- setting the CPU frequency to said target frequency;
- and
- updating the workload processing rate for said target frequency;
- wherein energy consumption is optimized under a performance slowdown constraint.
A computer-implemented method for optimizing energy consumption for an active computing event, comprising:
- (a) measuring a workload processing rate for a range of central processing unit (CPU) frequencies;
- (b) computing a run-time performance sensitivity factor relative to a change in CPU frequency;
- (c) computing a target frequency based on said run-time performance sensitivity factor;
- (d) wherein said target frequency comprises a minimum frequency having a predicted performance slowdown which does not exceed a maximum possible performance slowdown;
- (e) setting the CPU frequency to said target frequency;
- (f) updating the workload processing rate for said target frequency;
- and
- (g) repeating (b) through (f) until the active event is no longer active.
Dependencies are inferred from the claim text.