English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Most distributed-memory bulk-synchronous parallel programs in HPC assume thatcompute resources are available continuously and homogeneously across theallocated set of compute nodes. However, long one-off delays on individualprocesses can cause global disturbances, so-called idle waves, by ripplingthrough the system. This process is mainly governed by the communicationtopology of the underlying parallel code. This paper makes significantcontributions to the understanding of idle wave dynamics. We study thepropagation mechanisms of idle waves across the ranks of MPI-parallel programs.We present a validated analytic model for their propagation velocity withrespect to communication parameters and topology, with a special emphasis onsparse communication patterns. We study the interaction of idle waves with MPIcollectives and show that, depending on the implementation, a collective may betransparent to the wave. Finally we analyze two mechanisms of idle wave decay:topological decay, which is rooted in differences in communicationcharacteristics among parts of the system, and noise-induced decay, which iscaused by system or application noise. We show that noise-induced decay islargely independent of noise characteristics but depends only on the overallnoise power. An analytic expression for idle wave decay rate with respect tonoise power is derived. For model validation we use microbenchmarks and stencilalgorithms on three different supercomputing platforms.

Analytic Modeling of Idle Waves in Parallel Programs: Communication, Cluster Topology, and Noise Impact