Evaluating the Run-Time Performance of Kahn Process Network Implementation Techniques on Shared-Memory Multiprocessors

Vrba, Zeljko; Halvorsen, Paal; Griwodz, Carsten

doi:10.1109/cisis.2009.43

Cited by 6 publications

(12 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have measured that the cost of a single message send/receive operation combined with a context switch takes ∼ 0.8µs. Chapter 4 and section 5.2 extend the results published in [23].…”

Section: What Is the Effect Of Kpn Scheduling Policies On Applicationsupporting

confidence: 60%

Implementation and performance aspects of Kahn process networks

Vrba¹

2010

SIGMultimedia Rec.

Self Cite

View full text Add to dashboard Cite

Non quia difficilia sunt non audemus, sed quia non audemus, difficilia sunt. AbstractThe appearance of commodity multi-core processors, has spawned a wide interest in parallel programming, which is widely-regarded as more challenging than sequential programming. KPNs are a model of concurrency that relies exclusively on message passing, and that has some advantages over parallel programming tools in wide use today: simplicity, graphical representation, and determinism. Because of determinism, it is possible to reliably reproduce faults, an otherwise notoriously difficult problem with parallel programs. KPNs have gained acceptance in simulation and signal-processing communities. In this thesis, we investigate the applicability of KPNs to implementing general-purpose parallel computations for multi-core machines. In particular, we investigate 1) how KPNs can be used for modeling general-purpose problems; 2) how an efficient KPN run-time can be implemented; 3) what KPN scheduling strategies give good run-time performance. For these purposes, we have developed Nornir, an efficient run-time system for executing KPNs. With Nornir, we show that it is possible to develop a high-performance KPN run-time for multi-core machines. We experimentally demonstrate that problems expressed in the Kahn model resemble very much their sequential implementations, yet perform much better than when expressed in the MapReduce model, which has become widely-recognized as a simple parallel programming model. Lastly, we use Nornir to evaluate several load-balancing methods: static assignment, work-stealing, our improvement of work-stealing, and a method based on graph partitioning. The understanding brought by this evaluation is significant not only in the context of the Kahn model, but also in the more general context of load-balancing (potentially distributed) applications written in message-passing style.

show abstract

“…We have measured that the cost of a single message send/receive operation combined with a context switch takes ∼ 0.8µs. Chapter 4 and section 5.2 extend the results published in [23].…”

Section: What Is the Effect Of Kpn Scheduling Policies On Applicationsupporting

confidence: 60%

Implementation and performance aspects of Kahn process networks

Vrba¹

2010

SIGMultimedia Rec.

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, to obtain more general results, we have carefully designed the benchmark programs so that they execute correctly even when run-time deadlock detection and resolution is disabled. This is an otherwise key ingredient of a KPN run-time implementation [8], but it would make our observations less general as it would incur overheads not present in most applications.…”

Section: Methodsmentioning

confidence: 99%

“…In our earlier paper [8] we have noted that careful static assignment of processes to CPUs can match the performance of work-stealing on finely-granular parallel applications. Since static assignment is impractical for large process networks, we have also evaluated an automatic scheduling method based on graph partitioning by Devine et al [10], which balances the load across CPUs, and reduces the amount of inter-CPU communication as well as the cost of process migration.…”

Section: Introductionmentioning

confidence: 99%

“…Message-passing concurrency is an alternative to shared-state concurrency, and it has for a long time been used in distributed computing, and now also in modern parallel program frameworks like MapReduce [2], Oivos [3], and Dryad [4]. However, message passing frameworks also have an increasing importance on multi-core architectures, and such parallel program runtimes are being implemented and ported to single multi-core machines [5][6][7][8].…”

Section: Introductionmentioning

confidence: 99%

“…Both methods are implemented and tested in Nornir [8], which is our parallel processing runtime for executing parallel programs expressed as Kahn process networks [11].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Limits of Work-Stealing Scheduling

Vrba

Espeland

Halvorsen

et al. 2009

Job Scheduling Strategies for Parallel Processing

Self Cite

View full text Add to dashboard Cite

Abstract. The number of applications with many parallel cooperating processes is steadily increasing, and developing efficient runtimes for their execution is an important task. Several frameworks have been developed, such as MapReduce and Dryad, but developing scheduling mechanisms that take into account processing and communication requirements is hard. In this paper, we explore the limits of work stealing scheduler, which has empirically been shown to perform well, and evaluate loadbalancing based on graph partitioning as an orthogonal approach. All the algorithms are implemented in our Nornir runtime system, and our experiments on a multi-core workstation machine show that the main cause of performance degradation of work stealing is when very little processing time, which we quantify exactly, is performed per message. This is the type of workload in which graph partitioning has the potential to achieve better performance than work-stealing.

show abstract

The Nornir run-time system for parallel programs using Kahn process networks on multi-core machines—a flexible alternative to MapReduce

et al. 2010

Self Cite

View full text Add to dashboard Cite

Even though shared-memory concurrency is a paradigm frequently used for developing parallel applications on small-and middle-sized machines, experience has shown that it is hard to use. This is largely caused by synchronization primitives which are low-level, inherently non-deterministic, and, consequently, non-intuitive to use. In this paper, we present the Nornir run-time system. Nornir is comparable to well-known frameworks such as MapReduce and Dryad that are recognized for their efficiency and simplicity. Unlike these frameworks, Nornir also supports process structures containing branches and cycles. Nornir is based on the formalism of Kahn process networks, which is a shared-nothing, message-passing model of concurrency. We deem this model a simple and deterministic alternative to shared-memory concurrency. Experiments with real and synthetic benchmarks on up to 8 CPUs show that performance in most cases scales almost linearly with the number of CPUs, when not limited by data dependencies. We also show that the modeling flexibility allows Nornir to outperform its MapReduce counterparts using well-known benchmarks.

show abstract

Evaluating the Run-Time Performance of Kahn Process Network Implementation Techniques on Shared-Memory Multiprocessors

Abstract: Abstract

Cited by 6 publications

References 7 publications

Implementation and performance aspects of Kahn process networks

Implementation and performance aspects of Kahn process networks

Limits of Work-Stealing Scheduling

The Nornir run-time system for parallel programs using Kahn process networks on multi-core machines—a flexible alternative to MapReduce

Contact Info

Product

Resources

About