Efficient Scheduling of Nested Parallel Loops on Multi-Core Systems

Kejariwal, Arun; Nicolau, Alexandru; Veidenbaum, Alexander V.; Banerjee, Utpal; Polychronopoulos, Constantine D.

doi:10.1109/icpp.2009.19

Cited by 8 publications

(2 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach can be systematized by firstly, determining parallel inner do loops that can benefit from parallelism, i.e., loops that have a workload higher than the overhead involved by their parallel implementation [3], secondly, partitioning the iterations of the selected parallel loops among concurrent threads. Thirdly, privatizing auxiliary data that don't require global updates, i.e., reads and writes from multiple participating threads and, finally, protecting global shared data updates using locks.…”

Section: Data-sharing Approachmentioning

confidence: 99%

Spectral evolution simulation on leading multi-socket, multicore platforms

Tabik

Mimica

Plata

et al. 2011

2011 18th International Conference on High Performance Computing

View full text Add to dashboard Cite

Spectral evolution simulations based on the observed Very Long Baseline Interferometry (VLBI) radio-maps are of paramount importance to understand the nature of extragalactic objects in astrophysics. This work analyzes the performance and scaling of a spectral evolution algorithm on three leading multi-socket, multi-core architectures. We evaluate three parallel models with different levels of datasharing: a sharing approach, a privatizing approach and a hybrid approach. Our experiments show that the dataprivatizing model is reasonably efficient on medium scale multi-socket, multi-core systems (up to 48 cores) while regardless algorithmic and scheduling optimizations, sharing approach is unable to reach acceptable scalability on more than one socket. However, the hybrid model with a specific level of data-sharing gives the best scalability over all the considered multi-socket, multi-core systems.

show abstract

Section: Data-sharing Approachmentioning

confidence: 99%

Spectral evolution simulation on leading multi-socket, multicore platforms

Tabik

Mimica

Plata

et al. 2011

2011 18th International Conference on High Performance Computing

View full text Add to dashboard Cite

show abstract

“…Existing compiler and run-time systems tune parameters such as the degree of parallelism of data-parallel (DOALL) loops and block size of loops, either statically or dynamically, to match the execution environment [3,5,14,17,41,44]. These systems are limited to optimizing array-based programs with communicationfree data-parallelism, where the performance impact of those parameters can be relatively easily modeled.…”

Section: Introductionmentioning

confidence: 99%

Parcae

Raman

Zaks

Lee

et al. 2012

Proceedings of the 33rd ACM SIGPLAN Conference on Programming Language Design and Implementation

View full text Add to dashboard Cite

Workload, platform, and available resources constitute a parallel program's execution environment. Most parallelization efforts statically target an anticipated range of environments, but performance generally degrades outside that range. Existing approaches address this problem with dynamic tuning but do not optimize a multiprogrammed system holistically. Further, they either require manual programming effort or are limited to array-based data-parallel programs.This paper presents Parcae, a generally applicable automatic system for platform-wide dynamic tuning. Parcae includes (i) the Nona compiler, which creates flexible parallel programs whose tasks can be efficiently reconfigured during execution; (ii) the Decima monitor, which measures resource availability and system performance to detect change in the environment; and (iii) the Morta executor, which cuts short the life of executing tasks, replacing them with other functionally equivalent tasks better suited to the current environment. Parallel programs made flexible by Parcae outperform original parallel implementations in many interesting scenarios.

show abstract

Efficient Thread Labeling for On-the-fly Race Detection of Programs with Nested Parallelism

Jun

2011

Communications in Computer and Information Science

View full text Add to dashboard Cite

Efficient Scheduling of Nested Parallel Loops on Multi-Core Systems

Cited by 8 publications

References 18 publications

Spectral evolution simulation on leading multi-socket, multicore platforms

Spectral evolution simulation on leading multi-socket, multicore platforms

Parcae

Efficient Thread Labeling for On-the-fly Race Detection of Programs with Nested Parallelism

Contact Info

Product

Resources

About