Architectural Support for Handling Jitterin Shared Memory Based Parallel Applications

Chandran, Sandeep; Kallurkar, Prathmesh; Gupta, Parul; Sarangi, Smruti R.

doi:10.1109/tpds.2013.127

Cited by 3 publications

(1 citation statement)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Performance variability refers to the inherent "jitter" observed in performance values between multiple independent runs of an identical application. This variability can be of concern, particu-larly in large scale systems such as HPC and Cloud systems (Chandran et al 2014, Lofstead et al 2010, Schad, Dittrich, and Quiané-Ruiz 2010. Specifically for this work, throughput variability is considered with respect to various instances of the IOzone benchmark being run over a cluster of identical machines with identical system and application level parameters.…”

Section: Problem Summarymentioning

confidence: 99%

Predicting System Performance by Interpolation Using a High-Dimensional Delaunay Triangulation

2017

High Performance Computing Symposium (HPC 2018)

View full text Add to dashboard Cite

When interpolating computing system performance data, there are many input parameters that must be considered. Therefore, the chosen multivariate interpolation model must be capable of scaling to many dimensions. The Delaunay triangulation is a foundational technique, commonly used to perform piecewise linear interpolation in computer graphics, physics, civil engineering, and geography applications. It has been shown to produce a simplex based mesh with numerous favourable properties for interpolation. While computation of the two-and three-dimensional Delaunay triangulation is a well-studied problem, there are numerous technical limitations to the computability of a high-dimensional Delaunay triangulation. This paper proposes a new algorithm for computing interpolated values from the Delaunay triangulation without computing the complete triangulation. The proposed algorithm is shown to scale to over 50 dimensions. Data is presented demonstrating interpolation using the Delaunay triangulation in a real world high performance computing system problem.

show abstract

Section: Problem Summarymentioning

confidence: 99%