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

Smoothed Particle Hydrodynamics (SPH) is essential for modeling complexlarge-deformation problems across various applications, requiring significantcomputational power. A major portion of SPH computation time is dedicated tothe Nearest Neighboring Particle Search (NNPS) process. While advanced NNPSalgorithms have been developed to enhance SPH efficiency, the potentialefficiency gains from modern computation hardware remain underexplored. Thisstudy investigates the impact of GPU parallel architecture, low-precisioncomputing on GPUs, and GPU memory management on NNPS efficiency. Our approachemploys a GPU-accelerated mixed-precision SPH framework, utilizinglow-precision float-point 16 (FP16) for NNPS while maintaining high precisionfor other components. To ensure FP16 accuracy in NNPS, we introduce a RelativeCoordinated-based Link List (RCLL) algorithm, storing FP16 relative coordinatesof particles within background cells. Our testing results show threesignificant speedup rounds for CPU-based NNPS algorithms. The first comes fromparallel GPU computations, with up to a 1000x efficiency gain. The second isachieved through low-precision GPU computing, where the proposed FP16-basedRCLL algorithm offers a 1.5x efficiency improvement over the FP64-basedapproach on GPUs. By optimizing GPU memory bandwidth utilization, theefficiency of the FP16 RCLL algorithm can be further boosted by 2.7x, asdemonstrated in an example with 1 million particles. Our code is released athttps://github.com/pnnl/lpNNPS4SPH.

A GPU accelerated mixed-precision Smoothed Particle Hydrodynamics framework with cell-based relative coordinates