MG-FIM: A Multi-GPU Fast Iterative Method Using Adaptive Domain Decomposition

Hong, Sumin; Jang, Ganghee; Jeong, Won-Ki

doi:10.1137/21m1414644

Search citation statements

Order By: Relevance

Paper Sections

Select...

Numerical Considerations1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2023

Publication Types

Select...

Article1

Relationship

Self Cite0

Independent1

Authors

Journals

Cited by 1 publication

(1 citation statement)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…( 5). It relies on the well-developed velocity extension algorithm 19 , which has already been implemented with efficient algorithms 19,22,23 and has been parallelized and accelerated (e.g., on GPUs) 23,24 . These algorithms can be directly applied to the proposed method.…”

Section: Numerical Considerationsmentioning

confidence: 99%

Phase field modeling with large driving forces

Zhang,

Chadwick,

Chopp

et al. 2023

npj Comput Mater

View full text Add to dashboard Cite

There is growing interest in applying phase field methods as quantitative tools in materials discovery and development. However, large driving forces, common in many materials systems, lead to unstable phase field profiles, thus requiring fine spatial and temporal resolution. This demands more computational resources, limits the ability to simulate systems with a suitable size, and deteriorates the capability of quantitative prediction. Here, we develop a strategy to map the driving force to a constant perpendicular to the interface. Together with the third-order interpolation function, we find a stable phase field profile that is independent of the magnitude of the driving force. The power of this approach is illustrated using three models. We demonstrate that by using the driving force extension method, it is possible to employ a grid size orders of magnitude larger than traditional methods. This approach is general and should apply to many other phase field models.

show abstract