Energy-Aware Backfill Scheduling with DVFS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large-scale computing environments, existing scheduling techniques fail to efficiently balance computational speed with power consumption, leading to increased energy usage and wait times due to the lack of effective dynamic frequency and voltage scaling (DVFS) strategies, particularly in supercomputing facilities and data centers.
Innovation Solution
The method involves identifying job performance data, running simulations with various combinations of run-time overestimation values and processor adjustments, and applying dynamic voltage and frequency scaling (DVFS) to optimize the energy consumption and job delay product, thereby scheduling jobs to achieve energy-performance optimality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If dynamic frequency and voltage scaling (DVFS) is applied to reduce power consumption, then energy consumption is reduced, but computational speed deteriorates
Solution Approach 1:
The patent applies dynamic frequency and voltage scaling (DVFS) to make the processor operate at different frequency and voltage levels based on workload requirements. The scheduler dynamically adjusts processor speed during job execution, allowing the system to adapt between high performance and low power consumption states, thereby resolving the contradiction between energy efficiency and computational speed.
Solution Approach 2:
The patent changes the operational parameters (frequency and voltage) of the processor to optimize the trade-off between power consumption and performance. By adjusting these parameters dynamically based on job characteristics and system state, the system achieves energy-efficient operation without permanently sacrificing computational capability.
2Power
If processor frequency is decreased at fixed voltage to reduce power consumption, then power consumption is reduced, but computational speed is reduced
Solution Approach 1:
The system dynamically adjusts processor frequency based on actual workload demands and power constraints. Rather than using a fixed low frequency, the processor can scale up when computational tasks require higher performance and scale down during idle or less demanding periods, maintaining productivity while reducing average power consumption.
Solution Approach 2:
The scheduler periodically evaluates the system state and adjusts processor frequency accordingly. This periodic monitoring and adjustment allows the system to respond to changing workload conditions, ensuring that computational speed is maintained when necessary while achieving power savings during appropriate intervals.
3Use of energy by moving object
If DVFS settings are optimized to trade-off power expenditure against computational speed, then energy efficiency is improved, but scheduling complexity increases
Solution Approach 1:
The scheduling system incorporates feedback mechanisms that monitor job progress, power consumption, and system state. Based on this feedback, the scheduler adjusts DVFS settings in real-time to optimize energy efficiency. This feedback-driven approach automates the complex decision-making process, reducing the perceived scheduling complexity while maintaining high energy efficiency.
Solution Approach 2:
The system employs self-service mechanisms where the scheduler automatically selects and applies appropriate DVFS settings based on predefined policies and current system conditions. This automation eliminates the need for manual intervention in complex scheduling decisions, managing scheduling complexity internally while delivering optimized energy efficiency to the user.
4Productivity
If backfilling is used to improve scheduling performance, then overall scheduling efficiency is improved, but energy consumption management becomes more complex
Solution Approach 1:
The backfilling scheduler uses feedback from the system state to determine whether to accept backfill jobs and at what DVFS settings. This feedback mechanism allows the scheduler to balance the benefits of improved scheduling efficiency against the complexities of energy management by making informed, automated decisions about job placement and processor configuration.
Solution Approach 2:
The system performs preliminary evaluation of backfill opportunities and their impact on energy consumption before committing to scheduling decisions. By assessing potential backfill jobs in advance and predicting their energy implications, the scheduler can manage energy consumption proactively while maintaining high scheduling efficiency, reducing the complexity of real-time energy management.
Data Source
AI summary
An energy-aware backfill scheduling method combines overestimation of job run-times and processor adjustments, such as dynamic voltage and frequency scaling, to balance overall schedule performance and energy consumption. Accordingly, some scheduled jobs are executed in a manner reducing energy consumption. A computer-implemented method comprises identifying job performance data for a plurality of representative jobs and running a simulation of backfill-based job scheduling of the jobs at various combinations of run-time over-estimation values and processor adjustment values. The simulation generates data including energy consumption and job delay. The method further identifies one of the combinations of values that optimizes the mathematical product of an energy consumption parameter and a job delay parameter using the simulation generated data for the plurality of jobs. Jobs submitted to a processor are then scheduled using the identified combination of a run-time over-estimation value and a processor adjustment value.