2023
DOI: 10.1002/cpe.7509
|View full text |Cite
|
Sign up to set email alerts
|

Implicit propagation of directly addressed grids in lattice Boltzmann methods

Abstract: Lattice Boltzmann methods (LBM) are well suited to highly parallel computational fluid dynamics simulations due to their separability into a perfectly parallel collision step and a propagation step that only communicates within a local neighborhood. The implementation of the propagation step provides constraints for the maximum possible bandwidth-limited performance, memory layout and usage of vector instructions. This article revisits and extends the work on implicit propagation on directly addressed grids st… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
7
1

Relationship

2
6

Authors

Journals

citations
Cited by 11 publications
(3 citation statements)
references
References 37 publications
0
3
0
Order By: Relevance
“…through the space-time cylinder. The computing steps implemented in one collision are summarized in Algorithm 1 and the streaming is realized as a mere pointer shift [21]. The present work illustrates only the bulk solver.…”
Section: 3mentioning
confidence: 99%
“…through the space-time cylinder. The computing steps implemented in one collision are summarized in Algorithm 1 and the streaming is realized as a mere pointer shift [21]. The present work illustrates only the bulk solver.…”
Section: 3mentioning
confidence: 99%
“…While the underlying LBM library OpenLB supports various parallelization modes both on CPU and GPU targets [18,25], paint2sim explicitly only uses single-threaded, single-precision, non-vectorized execution for maximum device portability and as a heatmanagement trade-off. While OpenMP-based shared memory parallelization was possible, core heterogeneity and thread binding caused issues across the diverse set of test devices.…”
Section: Performancementioning
confidence: 99%
“…The successful implementation of the constructed algorithm depends on the memory access pattern. This topic is widely discussed in LBM implementations [9,53,54,[59][60][61]. The goal is not only the locality of data that are accessed in one sub-task, but also data alignment and coalescing.…”
Section: Data Structurementioning
confidence: 99%