Empirical Mode Decomposition for Modeling of Parallel Applications on Intel Xeon Phi Processors

Lawson, Gary; Sosonkina, Masha; Ezer, Tal; Shen, Yuzhong

doi:10.1109/ccgrid.2017.99

Cited by 3 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This process is repeated until the residual is found from which no other IMFs may obtained. One potential use for the residual trend is to construct a nonlinear model to relate power and time-to-solution, as proposed by Lawson et al (2017). Note that the total number of IMFs is an output of EMD and depends on the trace characteristics.…”

Section: Analysis Methodsmentioning

confidence: 99%

“…The intensity for the maximum number of cores is moderate, and for the minimum number of cores, the intensity is moderate-to-low; factoring time-to-solution with this difference, it is apparent that a moderate intensity coincides with the more optimal execution. It has been observed earlier (Lawson et al, 2017) that GAMESS is a memory-intensive application that achieves optimal performance with half of the maximum number of cores on Rulfo because of the limited L2 cache size (32 MB for 64 cores). Indeed, for GAMESS, a moderate intensity is seen in the 32-core trace (Figure 4(e)) while the plot for 63-core trace exhibits high intensity, notably between 50 and 200 s (Figure 4(b)).…”

Section: Comparisons Across Configurations and Applicationsmentioning

confidence: 98%

See 1 more Smart Citation

Applying EMD/HHT analysis to power traces of applications executed on systems with Intel Xeon Phi

Lawson

Sosonkina

Ezer

et al. 2017

The International Journal of High Performance Computing Applica

Self Cite

View full text Add to dashboard Cite

Power draw is a complex physical response to the workload of a given application on the hardware, which is difficult to model, in part, due to its variability. The empirical mode decomposition and Hilbert–Huang transform (EMD/HHT) is a method commonly applied to physical systems varying with time to analyze their complex behavior. In authors’ work, the EMD/HHT is considered for the first time to study power usage of high-performance applications. Here, this method is applied to the power measurement sequences (called here power traces) collected on three different computing platforms featuring two generations of Intel Xeon Phi, which are an attractive solution under the power budget constraints. The high-performance applications explored in this work are codesign molecular synamics and general atomic and molecular electronic structure system—which exhibit different power draw characteristics—to showcase strengths and limitations of the EMD/HHT analysis. Specifically, EMD/HHT measures intensity of an execution, which shows the concentration of power draw with respect to execution time and provides insights into performance bottlenecks. This article compares intensity among executions, noting on a relationship between intensity and execution characteristics, such as computation amount and data movement. In general, this article concludes that the EMD/HHT method is a viable tool to compare application power usage and performance over the entire execution and that it has much potential in selecting most appropriate execution configurations.

show abstract

Section: Analysis Methodsmentioning

confidence: 99%

Section: Comparisons Across Configurations and Applicationsmentioning

confidence: 98%