Transparent Resource Elasticity for Task-Based Cluster Environments with Work Stealing

Posner, Jonas; Fohry, Claudia

doi:10.1145/3458744.3473361

Cited by 10 publications

(5 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent work by Posner and Fohry with the APGAS runtime 19 made it possible to dynamically start and stop processes. This is of particular interest to us as PlhamJ could benefit from the capability to shrink and grow the number of processes it is running on as the workload of the simulation dynamically evolves over time.…”

Section: Discussionmentioning

confidence: 99%

Automatically balancing relocatable distributed collections

Finnerty

Kamada

Ohta

2023

Concurrency and Computation

View full text Add to dashboard Cite

SummaryIn previous work, we introduced a distributed collections library for the APGAS for Java programming model. This library makes it possible for programmers to develop complex distributed programs thanks to the many abstractions and computation patterns supported. In particular, programmers can fully and dynamically change the distribution of data entries through high‐level abstractions. However, the problem of balancing the load between processes remains, especially in cases where multiple processes may be concurrently executing on a single host, or when the performance of the hosts used differs. To address this issue and to relieve the burden of programming a load balancing strategy for a specific application, we created a dynamic load balancer integrated into our library. This load balancer operates within a specific context in a manner which does not interfere with the program legibility. Internally, we implement a scheme inspired by the lifeline‐based global load balancer scheme first introduced in 10. We evaluate the performance of our integrated load balancer on a small‐scale Beowulf cluster.

show abstract

Section: Discussionmentioning

confidence: 99%

Automatically balancing relocatable distributed collections

Finnerty

Kamada

Ohta

2023

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…This would reduce the waiting time and thus the turnaround time. A number of papers have demonstrated a reduction in both metrics by employing malleable jobs [6][7][8]. Another beneficial consideration for jobs is to take advantage of situations where there are free nodes and no jobs in the queue that require those nodes.…”

Section: Motivation and Backgroundmentioning

confidence: 99%

“…An efficient distribution of nodes makes it possible to assign to each job only the number of nodes that will give it maximum efficiency. In this way, it is possible to increase the throughput of the system [6][7][8]. On the other hand, it is possible to focus on another efficient distribution related to increasing the energy efficiency of each job [9][10][11].…”

Section: Motivation and Backgroundmentioning

confidence: 99%

See 1 more Smart Citation

Proteo: A Framework for the Generation and Evaluation of Malleable MPI Applications

Martín-Álvarez,

Aliaga,

Castillo

et al. 2024

Preprint

View full text Add to dashboard Cite

Applying malleability to HPC systems can increase their productivity without degrading or even improving the performance of running applications.This paper presents Proteo, a configurable framework that allows to design benchmarks to study the effect of malleability on a system, and also incorporates malleability into a real application. Proteo consists of two modules: SAM allows to emulate the computational behavior of iterative scientific MPI applications; and MaM is able to reconfigure a job during execution, adjusting the number of processes, redistributing data and resuming execution. An in-depth study of all the possibilities shows that Proteo is able to behave like a real malleable or non-malleable application in the range [0.85, 1.15]. Furthermore, the different methods defined in MaM for process management and data redistribution are analyzed, concluding that asynchronous malleability, where reconfiguration and application execution overlap, results in a 1.15x speedup.

show abstract

“…The application of malleability has shown an approximately 20% reduction of the makespan in Posner and Fohry (2021). Furthermore, Iserte et al (2020) demonstrated the ways to use it to reduce makespan by approximately 4x when combined with malleability techniques, whereas Iserte and Rojek (2019) showed its impact on energy efficiency improving approximately 2.4x in GPU-capable workloads.…”

Section: Introductionmentioning

confidence: 99%

Dynamic spawning of MPI processes applied to malleability

Martín-Álvarez

Estellés

Castillo

et al. 2023

The International Journal of High Performance Computing Applica

View full text Add to dashboard Cite

Malleability allows computing facilities to adapt their workloads through resource management systems to maximize the throughput of the facility and the efficiency of the executed jobs. This technique is based on reconfiguring a job to a different resource amount during execution and then continuing with it. One of the stages of malleability is the dynamic spawning of processes in execution time, where different decisions in this stage will affect how the next stage of data redistribution is performed, which is the most time-consuming stage. This paper describes different methods and strategies, defining eight different alternatives to spawn processes dynamically and indicates which one should be used depending on whether a strong or weak scaling application is being used. In addition, it is described for both types of applications which strategies benefit most the application performance or the system productivity. The results show that reducing the number of spawning processes by reusing the older ones can reduce reconfiguration time compared to the classical method by up to 2.6 times for expanding and up to 36 times for shrinking. Furthermore, the asynchronous strategy requires analysing the impact of oversubscription on application performance.

show abstract

Transparent Resource Elasticity for Task-Based Cluster Environments with Work Stealing

Cited by 10 publications

References 24 publications

Automatically balancing relocatable distributed collections

Automatically balancing relocatable distributed collections

Proteo: A Framework for the Generation and Evaluation of Malleable MPI Applications

Dynamic spawning of MPI processes applied to malleability

Contact Info

Product

Resources

About