Process Placement in Multicore Clusters:Algorithmic Issues and Practical Techniques

Jeannot, Emmanuel; Mercier, Guillaume; Tessier, François

doi:10.1109/tpds.2013.104

Cited by 91 publications

(116 citation statements)

References 31 publications

Supporting

Mentioning

116

Contrasting

Order By: Relevance

“…If k = 2 and the two graphs to be partitioned are the application graph and G p , the method is called dual recursive bipartitioning. Recently, schemes that model the processor graph as a tree have emerged [CLA12] in this algorithmic context and in similar ones [JMT13].…”

Section: Mapping Techniquesmentioning

confidence: 99%

Recent Advances in Graph Partitioning

Buluç

Meyerhenke

Safro

et al. 2016

Lecture Notes in Computer Science

413

300

View full text Add to dashboard Cite

show abstract

Section: Mapping Techniquesmentioning

confidence: 99%

Recent Advances in Graph Partitioning

Buluç

Meyerhenke

Safro

et al. 2016

Lecture Notes in Computer Science

413

300

View full text Add to dashboard Cite

show abstract

“…This information is used for placing tasks (usually threads or processes) in an affinity-aware way: matching inter-task affinities (tasks that synchronize/communicate a lot benefit from shorter distance between them) [2], [3]; placing tasks and their target data buffers together (on NUMA nodes close to cores and/or I/O devices that access them) [4].…”

Section: A Why Locality Is Importantmentioning

confidence: 99%

Memory Footprint of Locality Information on Many-Core Platforms

Goglin

2018

2018 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

View full text Add to dashboard Cite

Abstract-Exploiting the power of HPC platforms requires knowledge of their increasingly complex hardware topologies. Multiple components of the software stack, for instance MPI implementations or OpenMP runtimes, now perform their own topology discovery to find out the available cores and memory, and to better place tasks based on their affinities.We study in this article the impact of this topology discovery in terms of memory footprint. Storing locality information wastes an amount of physical memory that is becoming an issue on many-core platforms on the road to exascale.We demonstrate that this information may be factorized between processes by using a shared-memory region. Our analysis of the physical and virtual memories in supercomputing architectures shows that this shared region can be mapped at the same virtual address in all processes, hence dramatically simplifying the software implementation.Our implementation in hwloc and Open MPI shows a memory footprint that does not increase with the number of MPI ranks per node anymore. Moreover the job launch time is decreased by more than a factor of 2 on an Intel Knights Landing Xeon Phi and on a 96-core NUMA platform.

show abstract

“…To compute the allocation we use Algorithm 1 that is based on the TreeMatch algorithm [5]. We have adapted it in two ways to our needs.…”

Section: Topology-aware Add-on Of Orwl Tasksmentioning

confidence: 99%

Optimizing Locality by Topology-Aware Placement for a Task Based Programming Model

Gustedt

Jeannot

Mansouri

2016

2016 IEEE International Conference on Cluster Computing (CLUSTER)

Self Cite

View full text Add to dashboard Cite

Abstract-The ordered read-write lock model (ORWL) is a modern framework that proposes high level abstractions for the decomposition of an application and for the management of synchronizations and communications. The implementation of the model reaches high performances thanks to a decentralized event-based runtime. In this paper, we propose to enrich ORWL by proposing a topology-aware placement module that is based on the Hardware Locality framework, HWLOC. The aim is double. On one hand we increase the abstraction and the portability of the framework, and on the other hand we enhance the performance of the model's runtime. We propose a placement policy, that takes the characteristics of the application, of the runtime and of the architecture into account. We validate and compare our approach with the Livermore kernel23 benchmarks.

show abstract

Process Placement in Multicore Clusters:Algorithmic Issues and Practical Techniques

Cited by 91 publications

References 31 publications

Recent Advances in Graph Partitioning

Recent Advances in Graph Partitioning

Memory Footprint of Locality Information on Many-Core Platforms

Optimizing Locality by Topology-Aware Placement for a Task Based Programming Model

Contact Info

Product

Resources

About