Physics mesh tool speeds simulation setups on GPUs for AI models
PhysicsNeMo-Mesh: GPU-Native Meshes and Simulation Cases for Physics AI
Computational GeometryMachine Learning
Summary
Working with physical simulations is slow because the data describing shapes and conditions is large and complicated. The authors created PhysicsNeMo-Mesh, a tool that processes these shapes and simulation details much faster using graphics cards (GPUs) while keeping important information intact. This speedup helps different AI models share the same data more flexibly without expensive preparation steps each time. Their tool also keeps the structure needed for advanced AI designs that try to respect the physics of the problem.
What this means in practice
- •For machine learning engineers: Speed up physical simulation data preprocessing on GPUs to train and test multiple AI models efficiently with shared datasets.
- •For computational engineers: Use fast mesh processing inside simulation workflows to more accurately apply boundary conditions and reduce predictive errors in heat conduction problems.
Authors
Peter D. Sharpe, Corey Adams, Negin Sobhani, Sanjay Choudhry
Abstract
PhysicsNeMo-Mesh is the mesh module of the open-source PhysicsNeMo library. It provides mesh processing similar to PyVista, but on the GPU, in batches, and inside PyTorch's automatic differentiation. Like xarray for unstructured meshes, it attaches semantics to data: fields on points, cells, and whole meshes have names and shapes, and a simulation case keeps the complete PDE problem specification, including boundary conditions and their prescribed values. It addresses three problems that plague machine learning for physical simulation. First, GPU utilization is bottlenecked by I/O, because samples are very large relative to the compute per forward pass. Second, because traditional preprocessing is expensive, it is necessarily done once, during curation, rather than differently for each model and workflow at the point of use. Finally, advanced equivariant and mesh-native architectures require information that curation ordinarily discards. Fast GPU mesh processing makes it practical to preprocess at the point of use: on one GB200 GPU, mesh operations run $6.2\text{--}1{,}068\times$ faster than VTK with 72 threads, and loading a whole case is $8.0\text{--}88\times$ faster than reading its VTK files ($8.0\times$ when both read the same fields). Overlapped with training, this preprocessing costs $2.1\%$ of throughput for a $127\,\text{ms}$ Transolver step. So one dataset serves many models and workflows: in PhysicsNeMo's public recipe, one stored DrivAerML dataset feeds Transolver, GeoTransolver, and GLOBE. PhysicsNeMo-Mesh combines the speed of an intermediate array store with the flexibility of an archival format such as VTK. When shapes and wall conditions both vary, per-boundary conditions cut a heat-conduction error from $87\%$ to $6\text{--}9\%$. And by keeping field ranks and named boundaries, it retains the semantic structure that next-generation architectures will need.