# PhysicsNeMo-Mesh: GPU-Native Meshes and Simulation Cases for Physics AI

DevFeed: [PhysicsNeMo-Mesh: GPU-Native Meshes and Simulation Cases for Physics AI](<https://devfeed.tech/articles/physicsnemo-mesh-gpu-native-meshes-and-simulation-cases-for-physics-ai-108765.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.39072>)

Author: Peter D. Sharpe, Corey Adams, Negin Sobhani, Sanjay Choudhry

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [architectures](<https://devfeed.tech/tags/architectures.md>), [automatic](<https://devfeed.tech/tags/automatic.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [cases](<https://devfeed.tech/tags/cases.md>), [cells](<https://devfeed.tech/tags/cells.md>), [complete](<https://devfeed.tech/tags/complete.md>), [compute](<https://devfeed.tech/tags/compute.md>), [cs-cg](<https://devfeed.tech/tags/cs-cg.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [curation](<https://devfeed.tech/tags/curation.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [error](<https://devfeed.tech/tags/error.md>), [fast](<https://devfeed.tech/tags/fast.md>), [faster](<https://devfeed.tech/tags/faster.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [mesh](<https://devfeed.tech/tags/mesh.md>), [physics](<https://devfeed.tech/tags/physics.md>), [physics-ai](<https://devfeed.tech/tags/physics-ai.md>), [physics-comp-ph](<https://devfeed.tech/tags/physics-comp-ph.md>), [physicsnemo](<https://devfeed.tech/tags/physicsnemo.md>), [simulation](<https://devfeed.tech/tags/simulation.md>)

## AI overview

PhysicsNeMo-Mesh brings batched GPU mesh processing into PyTorch automatic differentiation while preserving named fields, their locations and shapes, and simulation cases with PDE specifications and boundary values. On one GB200 GPU, the abstract reports mesh operations 6.2-1,068x faster than VTK with 72 threads, and whole-case loading 8.0-88x faster than reading VTK files (8.0x when both read the same fields). With preprocessing overlapped with training, the reported throughput cost is 2.1% for a 127 ms Transolver step. This enables one stored DrivAerML dataset to feed Transolver, GeoTransolver, and GLOBE in a public recipe. In a heat-conduction setting where shapes and wall conditions vary, per-boundary conditions reduced error from 87% to 6-9%.

## Source excerpt

arXiv:2610.39072v1 Announce Type: cross 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.