Multi GPU Runtime Benchmarks for the H12 2D Urban Flood Model
This repository contains benchmark logs for a GPU-accelerated H12 2-D urban flood model. The model supports CPU-only and GPU-enabled execution, selectable via a compile-time option. Benchmarks were conducted at two study areas:• Portland, USA (9.0 km²) with high-resolution experiments at 1 m, 2 m and 4 m.• Downstream Han River basin, Republic of Korea (2,386 km²) with large-domain experiments at 10 m and 20 m.Each dataset records a one-hour simulation under rainfall forcing. Logs are organised by basin, spatial resolution and run configuration.
Institutions
- Kumoh National Institute of Technology (KR)
- Nvidia (United Kingdom) (GB)
- The University of Texas at Arlington (US)
- Nvidia (United States) (US)
- APEC Climate Center (KR)
- University of Sheffield (GB)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2025-10-07
- DOI
- https://doi.org/10.5281/zenodo.17284592
- Primary Topic
- Reproductive Physiology in Livestock
- Type
- article
- Field-Weighted Citation Impact
- 0.00