Thönnes D, Rüde U (2022)
Publication Language: English
Publication Type: Conference contribution, Abstract of a poster
Publication year: 2022
Event location: Basel, Switzerland
In this work, we present how code generation techniques significantly improve the performance of the computational kernels in the HyTeG framework. This HPC framework combines the performance and memory advantages of matrix-free multigrid solvers with the flexibility of unstructured meshes. With the use of the pystencils code generation toolbox, the original abstract C++ kernels are replaced with highly optimized loop nests. The performance of one of those kernels (the matrix-vector multiplication) is thoroughly analyzed using the Execution-Cache-Memory (ECM) performance model. These predictions are validated by measuring execution on the SuperMUC-NG supercomputer. Overall, the results agree with the predicted performance, and the discrepancies are discussed. Additionally, we conduct a node-level scaling study which shows the expected behavior for a memory-bound compute kernel.
APA:
Thönnes, D., & Rüde, U. (2022). Performance Modeling of Generated Stencil Kernels within the HyTeG Framework. Poster presentation at Platform for Advanced Scientific Computing (PASC) Conference, Basel, Switzerland.
MLA:
Thönnes, Dominik, and Ulrich Rüde. "Performance Modeling of Generated Stencil Kernels within the HyTeG Framework." Presented at Platform for Advanced Scientific Computing (PASC) Conference, Basel, Switzerland 2022.
BibTeX: Download