Metraipolis › archive › dc_5c721ea12ca2
Accelerating the Solving of Many Tiny General Linear Systems on GPUs: Application to Constitutive Laws — arXiv:2609.15217
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.15217
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T23:17:40.888Z
- SHA-256 of the archived text
b5b394ad32f3a019fc5ceff96f867eaae87a12f64d7815955da70041424dc6fb- Stable URL of this page
https://metraipolis.com/documento/dc_5c721ea12ca2
How to cite
Accelerating the Solving of Many Tiny General Linear Systems on GPUs: Application to Constitutive Laws — arXiv:2609.15217 — arXiv (en), https://arxiv.org/abs/2609.15217 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5c721ea12ca2 (Metraipolis archive dc_5c721ea12ca2, sha256 b5b394ad32f3a019fc5ceff96f867eaae87a12f64d7815955da70041424dc6fb)
BibTeX
@misc{dc5c721ea12ca2,
title = {Accelerating the Solving of Many Tiny General Linear Systems on GPUs: Application to Constitutive Laws — arXiv:2609.15217},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.15217},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5c721ea12ca2; sha256 b5b394ad32f3a019fc5ceff96f867eaae87a12f64d7815955da70041424dc6fb},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.15217. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Many applications require solving large numbers of independent linear systems on GPUs. While this need is well addressed for small to large systems, tiny ones, understood here as systems of dimension below 32, remain challenging. This is especially relevant in constitutive law evaluation, where millions of integration points are handled independently, and where each constitutive update generally relies on a Newton iterative method. Each iteration then requires the double-precision solution of a tiny general square linear system using LU factorization with partial pivoting (LUpp). The present study was conducted within a closed-source prototype, which serves as a demonstrator for porting to NVIDIA GPUs constitutive law evaluations currently provided on CPUs by TFEL/MFront, an open-source code generation tool for material knowledge. We compare several double-precision LUpp solvers, including implementations from GPU linear algebra libraries as well as custom-designed CUDA kernels. The comparison is performed first on large batches of standalone linear systems, and then within the full constitutive-law evaluation workflow, where each integration point requires a sequence of distinct linear systems, one per iteration of its own Newton loop. The study shows that the best LUpp solving strategy strongly depends on several factors including system size and application context. We discuss several key aspects, including register pressure, occupancy, the ability to invoke the LUpp solver directly from device code, and whether assigning several threads to each system is the most efficient strategy. Experiments on an NVIDIA H100 GPU show that the specialized LUpp solvers proposed in this work can outperform existing state-of-the-art approaches for this class of workloads, with speedups of up to 6.5x over cuSolverDx, and up to 17.7x over MAGMA. Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-15 in cs.DC). Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC). Pagina del lavoro: https://arxiv.org/abs/2609.15217 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T22:54:12.988Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).