Energy-aware job scheduling for cluster environments

• US 8,612,984 B2
• Filed: 11/01/2010
• Issued: 12/17/2013
• Est. Priority Date: 04/28/2010
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bell, Robert H. Jr., Brochard, Luigi, DeSota, Donald R., Panda, Rajendra D., Thomas, Francois
• Primary Examiner(s): Truong, Camquy
• Application Number:
  US12/917,421
• Publication Number:
  US 20110271283A1
• Time in Patent Office: 1,142 Days
• Field of Search: None
• US Class Current: 718/102
• CPC Class Codes: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A job scheduler can select a processor core operating frequency for a node in a cluster to perform a job based on energy usage and performance data. After a job request is received, an energy aware job scheduler accesses data that specifies energy usage and job performance metrics that correspond to the requested job and a plurality of processor core operating frequencies. A first of the plurality of processor core operating frequencies is selected that satisfies an energy usage criterion for performing the job based, at least in part, on the data that specifies energy usage and job performance metrics that correspond to the job. The job is assigned to be performed by a node in the cluster at the selected first of the plurality of processor core operating frequencies.

Claims

1. A computer program product for energy-aware job scheduling, the computer program product comprising:

a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising;

computer readable program code configured to,

access data that specifies energy usage and job performance metrics that correspond to a job and a plurality of processor core operating frequencies, wherein the job has been requested of a cluster;

select a first of the plurality of processor core operating frequencies that satisfies an energy usage criterion for performing the job based, at least in part, on the data that specifies energy usage and job performance metrics that correspond to the job,

wherein the plurality of processor core operating frequencies comprises a default processor core operating frequency, and wherein the first of the plurality of processor core operating frequencies is lower than the default processor core operating frequency;

determine coefficients of an energy model based on the performance metrics and energy usage;

estimate energy usage and performance metrics for the job at the first of the plurality of processor core operating frequencies based on the coefficients and the energy model; and

store the estimated energy usage and performance metrics for the job at the first of the plurality of processor core operating frequencies;

associate the stored estimated energy usage and performance metrics with the job and with the first of the plurality of processor core operating frequencies; and

assign the job to be performed by a node in the cluster at the selected first of the plurality of processor core operating frequencies.

2. The computer program product of claim 1, wherein the computer readable program code being configured to access the data comprises the computer readable program code being configured to determine that the job has previously been performed or that a similar job has been performed.

3. The computer program product of claim 2, wherein the energy usage criterion for performing the job indicates at least one of a maximum amount of energy to be used to perform the job and an energy conservation goal.

4. The computer program product of claim 2, wherein the computer readable program code being configured to assign the job to be performed by the node in the cluster at the first of the plurality of processor core operating frequencies comprises the computer readable program code being configured to:

set at least a first processor core of the node to run at the first of the plurality of processor core operating frequencies.

5. The computer program product of claim 1, wherein the computer readable program code is further configured to

prior to said selecting the first of the plurality of processor core operating frequencies that satisfies the energy usage criterion for performing the job, collect energy usage and performance metrics for the job performed in the cluster at the default processor core operating frequency.

6. An apparatus comprising:

a processor;

a network interface; and

an energy aware job scheduler operable to,

evaluate energy usage and job performance metrics against an energy usage criterion for a job, wherein the energy usage and job performance metrics correspond to the job at different ones of a plurality of processor core operating frequencies of a node in a cluster;

select a first of the plurality of processor core operating frequencies based, at least in part, on evaluation of the energy usage and job performance metrics against the energy usage criterion for the job,

wherein the plurality of processor core operating frequencies comprises a default processor core operating frequency, and wherein the first of the plurality of processor core operating frequencies is lower than the default processor core operating frequency;

determine coefficients of an energy model based on the performance metrics and energy usage;

estimate energy usage and performance metrics for the job at the first of the plurality of processor core operating frequencies based on the coefficients and the energy model; and

store the estimated energy usage and performance metrics for the job at the first of the plurality of processor core operating frequencies;

associate the stored estimated energy usage and performance metrics with the job and with the first of the plurality of processor core operating frequencies; and

assign the job to the node for performing the job at the first of the plurality of processor core operating frequencies.

7. The apparatus of claim 6, wherein the energy aware job scheduler is further operable to determine that the job or a similar job has previously been performed.

8. The apparatus of claim 6, wherein the energy aware job scheduler being operable to assign the job to the node for performing the job at the first of the plurality of processor core operating frequencies comprises the energy aware job scheduler being operable to set a core of the node to run at the first of the plurality of processor core operating frequencies.

9. The apparatus of claim 6, wherein the energy usage criterion indicates at least one of a maximum amount of energy to be used to perform the job and an energy conservation goal.