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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Realistic simulations in engineering or in the materials sciences can consumeenormous computing resources and thus require the use of massively parallelsupercomputers. The probability of a failure increases both with the runtimeand with the number of system components. For future exascale systems it istherefore considered critical that strategies are developed to make softwareresilient against failures. In this article, we present a scalable,distributed, diskless, and resilient checkpointing scheme that can create andrecover snapshots of a partitioned simulation domain. We demonstrate theefficiency and scalability of the checkpoint strategy for simulations with upto $40$ billion computational cells executing on more than $400$ billionfloating point values. A checkpoint creation is shown to require only a fewseconds and the new checkpointing scheme scales almost perfectly up to morethan $260\,000$ ($2^{18}$) processes. To recover from a diskless checkpointduring runtime, we realize the recovery algorithms using ULFM MPI. Thecheckpointing mechanism is fully integrated in a state-of-the-arthigh-performance multi-physics simulation framework. We demonstrate theefficiency and robustness of the method with a realistic phase-field simulationoriginating in the material sciences and with a lattice Boltzmann methodimplementation.

A Scalable and Extensible Checkpointing Scheme for Massively Parallel Simulations