Data Partitioning on Multicore and Multi-GPU Platforms Using Functional Performance Models

Zhong, Ziming; Рычков, В. Н.; Lastovetsky, Alexey

doi:10.1109/tc.2014.2375202

Cited by 51 publications

(31 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It has been shown [29,28] that in modern multicore, manycore and hybrid platforms, where processing elements are coupled and share resources, the speed of one group of elements may depend on the load of others due to resource contention. Therefore, the groups cannot be considered as independent processing units and their speed cannot be measured separately and independently.…”

Section: Applying Model-based Partitioning Algorithm To Mpdata Decompmentioning

confidence: 99%

“…Therefore, the groups cannot be considered as independent processing units and their speed cannot be measured separately and independently. In this work, we use the performance measurement method proposed in [28] . According to this method, the performance of the four teams of cores is measured simultaneously rather than separately, thereby taking into account resource contention.…”

Section: Applying Model-based Partitioning Algorithm To Mpdata Decompmentioning

confidence: 99%

See 1 more Smart Citation

Model-Based Optimization of EULAG Kernel on Intel Xeon Phi Through Load Imbalancing

Lastovetsky

Szustak

Wyrzykowski

2017

IEEE Trans. Parallel Distrib. Syst.

Self Cite

View full text Add to dashboard Cite

Load balancing is a widely accepted technique for performance optimization of scientific applications on parallel architectures. Indeed, balanced applications do not waste processor cycles on waiting at points of synchronization and data exchange, maximizing this way the utilization of processors. In this paper, we challenge the universality of the load-balancing approach to optimization of the performance of parallel applications. First, we formulate conditions that should be satisfied by the performance profile of an application in order for the application to achieve its best performance via load balancing. Then we use a real-life scientific application, MPDATA, to demonstrate that its performance profile on a modern parallel architecture, Intel Xeon Phi, significantly deviates from these conditions. Based on this observation, we propose a method of performance optimization of scientific applications through load imbalancing. We also propose an algorithm that finds the optimal, possibly imbalanced, configuration of a data parallel application on a set of homogeneous processors. This algorithm uses functional performance models of the application to find the partitioning that minimizes its computation time but not necessarily balances the load of the processors. We show how to apply this algorithm to optimization of MPDATA on Intel Xeon Phi. Experimental results demonstrate that the performance of this carefully optimized load-balanced application can be further improved by 15% using the proposed load-imbalancing optimization.

show abstract

Section: Applying Model-based Partitioning Algorithm To Mpdata Decompmentioning

confidence: 99%

Section: Applying Model-based Partitioning Algorithm To Mpdata Decompmentioning

confidence: 99%

Model-Based Optimization of EULAG Kernel on Intel Xeon Phi Through Load Imbalancing

Lastovetsky

Szustak

Wyrzykowski

2017

IEEE Trans. Parallel Distrib. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In [8] the distribution is obtained with automatic code modifications. Zhong et al [9] use performance models. These works fail to address the importance of adaptability to irregular loads.…”

Section: Related Workmentioning

confidence: 99%

Extending OmpSs for OpenCL Kernel Co-Execution in Heterogeneous Systems

Pérez

Stafford

Bosque

et al. 2017

2017 29th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

View full text Add to dashboard Cite

Abstract-Heterogeneous systems have a very high potential performance but present difficulties in their programming. OmpSs is a well known framework for task based parallel applications, which is an interesting tool to simplify the programming of these systems. However, it does not support the co-execution of a single OpenCL kernel instance on several compute devices. To overcome this limitation, this paper presents an extension of the OmpSs framework that solves two main objectives: the automatic division of datasets among several devices and the management of their memory address spaces. To adapt to different kinds of applications, the data division can be performed by the novel HGuided load balancing algorithm or by the well known Static and Dynamic. All this is accomplished with negligible impact on the programming. Experimental results reveal that there is always one load balancing algorithm that improves the performance and energy consumption of the system.

show abstract

“…This problem was recently addressed in [46] [47] [48]. In this work, the authors do not study how to develop computational kernels for individual computing devices used in hybrid heterogeneous platforms, such as multicore CPUs or GPUs.…”

Section: Optimization Of Parallel Applications On Hybrid Multicore Anmentioning

confidence: 99%

Heterogeneous parallel computing: from clusters of workstations to hierarchical hybrid platforms

Lastovetsky

2014

JSFI

View full text Add to dashboard Cite

show abstract

Data Partitioning on Multicore and Multi-GPU Platforms Using Functional Performance Models

Cited by 51 publications

References 27 publications

Model-Based Optimization of EULAG Kernel on Intel Xeon Phi Through Load Imbalancing

Model-Based Optimization of EULAG Kernel on Intel Xeon Phi Through Load Imbalancing

Extending OmpSs for OpenCL Kernel Co-Execution in Heterogeneous Systems

Heterogeneous parallel computing: from clusters of workstations to hierarchical hybrid platforms

Contact Info

Product

Resources

About