Affinity Driven Distributed Scheduling Algorithm for Parallel Computations

Narang, Ankur; Srivastava, Abhinav; Kumar, Naga Praveen; Shyamasundar, R. K.

doi:10.1007/978-3-642-17679-1_15

Cited by 2 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While both the above efforts are for a single place, we consider multiple places and provide potential function based analysis to derive expected and probabilistic lower and upper bounds on time and message complexity for our affinity driven algorithm. [11] presents an affinity distributed scheduling algorithm with high level theoretical results and experimental analysis limited to Intel shared memory NUMA architecture. This work is an extension of [11] with detailed theoretical analysis and also extensive experimental analysis on multi-core clusters including Intel multi-core cluster as well as Blue Gene/P with upto 256 places.…”

Section: Related Workmentioning

confidence: 99%

Performance driven multi-objective distributed scheduling for parallel computations

Narang

Srivastava

Katta

et al. 2011

SIGOPS Oper. Syst. Rev.

Self Cite

View full text Add to dashboard Cite

With the advent of many-core architectures and strong need for Petascale (and Exascale) performance in scientific domains and industry analytics, efficient scheduling of parallel computations for higher productivity and performance has become very important. Further, movement of massive amounts (Terabytes to Petabytes) of data is very expensive, which necessitates affinity driven computations. Therefore, distributed scheduling of parallel computations on multiple places 1 needs to optimize multiple performance objectives: follow affinity maximally and ensure efficient space, time and message complexity. Simultaneous consideration of these objectives makes distributed scheduling a particularly challenging problem. In addition, parallel computations have data dependent execution patterns which requires online scheduling to effectively optimize the computation orchestration as it unfolds.This paper presents an online algorithm for affinity driven distributed scheduling of multi-place 2 parallel computations. To optimize multiple performance objectives simultaneously, our algorithm uses a low time and message complexity mechanism for ensuring affinity and a randomized work-stealing mechanism within places for load balancing. Theoretical analysis of the expected and probabilistic lower and upper bounds on time and message complexity of this algorithm has been provided. On multi-core clusters such as Blue Gene/P (MPP architecture) and Intel multicore cluster, we demonstrate performance close to the custom MPI+Pthreads code. Further, strong, weak and data (increasing input data size) scalability have been demonstrated on multi-core clusters. Using well known benchmarks, we demonstrate 16% to 30% performance gain as compared to Cilk [6] on multi-core Intel Xeon 5570 (NUMA) architecture. Detailed experimental analysis illustrates efficient space (main memory) utilization as well. To the best of 1 place is a group of processors with shared memory 2 multi-place refers to a group of places. For example, with each place as an SMP(Symmetric MultiProcessor), multi-place refers to cluster of SMPs our knowledge, this is the first time multi-objective affinity driven distributed scheduling algorithm has been designed, theoretically analyzed and experimentally evaluated in a multi-place setup for multi-core cluster architectures.

show abstract

Section: Related Workmentioning

confidence: 99%

Performance driven multi-objective distributed scheduling for parallel computations

Narang

Srivastava

Katta

et al. 2011

SIGOPS Oper. Syst. Rev.

Self Cite

View full text Add to dashboard Cite

show abstract

Performance driven distributed scheduling of parallel hybrid computations

Narang

Shyamasundar

2011

Theoretical Computer Science

Self Cite

View full text Add to dashboard Cite

a b s t r a c tExascale computing is fast becoming a mainstream research area. In order to realize exascale performance, it is necessary to have efficient scheduling of large parallel computations with scalable performance on a large number of cores/processors. The scheduler needs to execute in a pure distributed and online fashion, should follow affinity inherent in the computation and must have low time and message complexity. Further, it should also avoid physical deadlocks due to bounded resources including space/memory per core. Simultaneous consideration of these factors makes affinity driven distributed scheduling particularly challenging. We attempt to address this challenge for hybrid parallel computations which contain tasks that have pre-specified affinity to a place and also tasks that can be mapped to any place in the system. Specifically, we address two scheduling problems of the type P m |M j , prec|C max . This paper presents online distributed scheduling algorithms for hybrid parallel computations assuming both unconstrained and bounded space per place. We also present the time and message complexity for distributed scheduling of hybrid computations. To the best of our knowledge, this is the first time that distributed scheduling algorithms for hybrid parallel computations have been presented and analyzed for time and message bounds under both unconstrained space and bounded space.

show abstract

Affinity Driven Distributed Scheduling Algorithm for Parallel Computations

Cited by 2 publications

References 8 publications

Performance driven multi-objective distributed scheduling for parallel computations

Performance driven multi-objective distributed scheduling for parallel computations

Performance driven distributed scheduling of parallel hybrid computations

Contact Info

Product

Resources

About