A versatile FEM framework with native GPU scalability via globally-applied AD
Mohit Pundir •, Flavio Lorez •, David S. Kammer • · 2026
Read A versatile FEM framework with native GPU scalability via globally-applied AD by Mohit Pundir •, Flavio Lorez •, David S. Kammer • (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-12365.txt.
Description
Energy-based finite-element methods describe complex mechanics from one global energy functional, whose derivatives give the governing equations. Mohit Pundir led this work with two co-authors, tackling a trade-off in automatic differentiation. Global differentiation is expressive but hard to scale, while local differentiation scales well but limits the physics and couplings a frame can express easily.
Their framework, tatva, defines physics as a single global functional and differentiates it globally to yield residual and tangent operators. Jacobian-vector products power matrix-free solvers, and colouring-based sparse differentiation builds sparse tangent stiffness matrices when needed. The design scales linearly on GPUs, handles millions of degrees of freedom without running out of memory, and covers constraints, mixed-dimensional coupling, and neural networks.
Catalog record
| CREATOR | Mohit Pundir •, Flavio Lorez •, David S. Kammer • |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 94 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-12365