Performance Evaluation of Heterogeneous GPU Programming Frameworks for Hemodynamic Simulations

Conference · Wed Nov 01 00:00:00 EDT 2023

DOI:https://doi.org/10.1145/3624062.3624188· OSTI ID:2224199

Martin, Aristotle ^[1]; Liu, Geng ^[2]; Ladd, William ^[1]; Lee, Seyong  ^[3]; Gounley, John ^[3]; Vetter, Jeffrey  ^[3]; Patel, Saumil ^[4]; Rizzi, Silvio ^[4]; Mateevitsi, Victor A. ^[2]; Insley, Joseph ^[4]; Randles, Amanda ^[5]

---

1. Duke University
2. Argonne National Laboratory
3. ORNL
4. Argonne National Laboratory (ANL)
5. Duke university Duhram, NC

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, fluid dynamics 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 along with a custom performance model to inform the results and identify additional optimization strategies. At scale performance of the programming model implementations are evaluated on pre-production GPU node architectures for Frontier and Aurora, as well as on current NVIDIA device-based systems Summit and Polaris. Real-world workloads representing 3D blood flow calculations in complex vasculature are assessed. Our analysis highlights critical trade-offs between code performance, portability, and development time.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-00OR22725
• OSTI ID: 2224199
• Resource Relation: Conference: 2023 International Workshop on Performance, Portability & Productivity in HPC - Denver, Colorado, United States of America - 11/12/2023 10:00:00 AM-11/17/2023 10:00:00 AM
• Country of Publication: United States
• Language: English