Obtaining dynamic scheduling policies with simulation and machine learning

Carastan-Santos, Danilo; Camargo, Raphael Y. de

doi:10.1145/3126908.3126955

Cited by 27 publications

(45 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Throughout this paper we use the bounded slowdown (BSLD) metric as it is accepted as one of the most popular one metrics to measure the performance of scheduling heuristics [7], [8]. The BSLD of a job j is defined as follows:…”

Section: Metricmentioning

confidence: 99%

“…Perhaps the most comparable works to the one presented in this paper are [14] and [8]. In [14], the authors developed DeepRM, a multi-resource cluster scheduler that uses deep reinforcement learning to solve the problem of packing with multiple resource demands.…”

Section: Related Workmentioning

confidence: 99%

“…In [14], the authors developed DeepRM, a multi-resource cluster scheduler that uses deep reinforcement learning to solve the problem of packing with multiple resource demands. In [8] Carastan-Santos and Camargo used synthetic workloads to create general heuristics that improve the slowdown metric. They combined the basic job characteristics in a non-linear function and they used linear regression to devise new heuristics.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Adapting Batch Scheduling to Workload Characteristics: What Can We Expect From Online Learning?

Legrand

Trustram

Zrigui

2019

2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

View full text Add to dashboard Cite

Despite the impressive growth and size of super-computers, the computational power they provide still cannot match the demand. Efficient and fair resource allocation is a critical task. Super-computers use Resource and Job Management Systems to schedule applications, which is generally done by relying on generic index policies such as First Come First Served and Shortest Processing time First in combination with Backfilling strategies. Unfortunately, such generic policies often fail to exploit specific characteristics of real workloads. In this work, we focus on improving the performance of online schedulers. We study mixed policies, which are created by combining multiple job characteristics in a weighted linear expression, as opposed to classical pure policies which use only a single characteristic. This larger class of scheduling policies aims at providing more flexibility and adaptability. We use space coverage and black-box optimization techniques to explore this new space of mixed policies and we study how can they adapt to the changes in the workload. We perform an extensive experimental campaign through which we show that (1) even the best pure policy is far from optimal and that (2) using a carefully tuned mixed policy would allow to significantly improve the performance of the system. (3) We also provide empirical evidence that there is no one size fits all policy, by showing that the rapid workload evolution seems to prevent classical online learning algorithms from being effective.

show abstract

Section: Metricmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Adapting Batch Scheduling to Workload Characteristics: What Can We Expect From Online Learning?

Legrand

Trustram

Zrigui

2019

2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

View full text Add to dashboard Cite

show abstract

“…The use of machine learning opens up the opportunity to move from fixed heuristic to dynamic policies for scheduling workloads. Carastan-Santos et al [10] provide one such work, demonstrating notable improvements in task slowdown against existing scheduling approaches.…”

Section: A Scheduling Decisionsmentioning

confidence: 99%

“…Figure 5 depicts a simple RF of three trees derived from three features (c -cluster number, m -machine number, dday of week) to predict the idle time. To predict the idle time for (c=1, m=1, d=1) each tree evaluates a prediction (10,10,11) with the value 10 returned.…”

Section: B Random Forestmentioning

confidence: 99%

Using Machine Learning to reduce the energy wasted in Volunteer Computing Environments

McGough

Forshaw

Brennan

et al. 2018

2018 Ninth International Green and Sustainable Computing Conference (IGSC)

View full text Add to dashboard Cite

High Throughput Computing (HTC) provides a convenient mechanism for running thousands of tasks. Many HTC systems exploit computers which are provisioned for other purposes by utilising their idle time -volunteer computing. This has great advantages as it gives access to vast quantities of computational power for little or no cost. The downside is that running tasks are sacrificed if the computer is needed for its primary use. Normally terminating the task which must be restarted on a different computer -leading to wasted energy and an increase in task completion time. We demonstrate, through the use of simulation, how we can reduce this wasted energy by targeting tasks at computers less likely to be needed for primary use, predicting this idle time through machine learning. By combining two machine learning approaches, namely Random Forest and MultiLayer Perceptron, we save 51.4% of the energy without significantly affecting the time to complete tasks.

show abstract

A HPC Co-scheduler with Reinforcement Learning

Souza

Pelckmans

Tordsson

2021

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

Although High Performance Computing (HPC) users understand basic resource requirements such as the number of CPUs and memory limits, internal infrastructural utilization data is exclusively leveraged by cluster operators, who use it to configure batch schedulers. This task is challenging and increasingly complex due to ever larger cluster scales and heterogeneity of modern scientific workflows. As a result, HPC systems achieve low utilization with long job completion times (makespans). To tackle these challenges, we propose a co-scheduling algorithm based on an adaptive reinforcement learning algorithm, where application profiling is combined with cluster monitoring. The resulting cluster scheduler matches resource utilization to application performance in a fine-grained manner (i.e., operating system level). As opposed to nominal allocations, we apply decision trees to model applications' actual resource usage, which are used to estimate how much resource capacity from one allocation can be co-allocated to additional applications. Our algorithm learns from incorrect co-scheduling decisions and adapts from changing environment conditions, and evaluates when such changes cause resource contention that impacts quality of service metrics such as jobs slowdowns. We integrate our algorithm in an HPC resource manager that combines Slurm and Mesos for job scheduling and co-allocation, respectively. Our experimental evaluation performed in a dedicated cluster executing a mix of four real different scientific workflows demonstrates improvements on cluster utilization of up to 51% even in high load scenarios, with 55% average queue makespan reductions under low loads.

show abstract

Obtaining dynamic scheduling policies with simulation and machine learning

Cited by 27 publications

References 26 publications

Adapting Batch Scheduling to Workload Characteristics: What Can We Expect From Online Learning?

Adapting Batch Scheduling to Workload Characteristics: What Can We Expect From Online Learning?

Using Machine Learning to reduce the energy wasted in Volunteer Computing Environments

A HPC Co-scheduler with Reinforcement Learning

Contact Info

Product

Resources

About