- Preparing for the deployment of large scientific and engineering codes on upcoming exascale systems with GPU-dense nodes is made challenging by the unprecedented diversity of device architectures and heterogeneous programming models. In this work, we evaluate the process of porting a massively parallel, multi-physics code written in CUDA to SYCL, HIP, and Kokkos with a range of backends, using a combination of automated tools and manual tuning. We use a proxy application alongside a custom performance model to inform results and identify additional optimization strategies. At scale performance of the programming model variants is evaluated on pre-production GPU node architectures for Frontier and Aurora, as well ... [Read More]
- Total Size
- 4 files (8 MB)
- Data Citation
- Martin, A., Liu, G., Ladd, W., Lee, S., Gounley, J., Vetter, J., Patel, S., Rizzi, S., Mateevitsi, V., Insley, J., Randles, A. (2023). Data and scripts from: Performance evaluation of heterogenous GPU programming frameworks for hemodynamic simulations. Duke Research Data Repository. https://doi.org/10.7924/r45t3t64k
- Creator
- DOI
- 10.7924/r45t3t64k
- Publication Date
- April 21, 2023
- ARK
- ark:/87924/r45t3t64k
- Affiliation
- Publisher
- Type
- Contact
- Aristotle Martin, aristotle.martin@duke.edu, 919-308-5702, https://orcid.org/0000-0002-8704-764X
- Title
- Data and scripts from: Performance evaluation of heterogenous GPU programming frameworks for hemodynamic simulations
- Repository
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
graphdat.zip | 2023-04-21 | Download | |
|
|
perfmodel.zip | 2023-04-21 | Download | |
|
|
proxyapps.zip | 2023-04-21 | Download | |
|
README.md | 2023-04-21 | Download |