Triplet Pruning and Spatial Decomposition in M3GNet: Accuracy and Memory Trade-offs in Scaling Graph Neural Network Potentials
- 1Applied Artificial Intelligence Institute
- 2Deakin University
Abstract
Machine-learning interatomic potentials of several functional forms now support large-scale atomistic simulation.Within this landscape, graph neural network (GNN) potentials such as M3GNet offer transferable message-passing models, but their application to very large systems remains limited by graph construction, message passing, and differentiable force-evaluation costs.In M3GNet, the explicit angular pathway is represented through line-graph triplets, whose construction can dominate inference cost.Here we examine the scalability of M3GNet molecular dynamics via inference-time triplet pruning.We first show that an apparent insensitivity to triplet removal in inspected MatGL implementations originates from an incorrect parent-bond index mapping; after correction, removing triplets substantially increases static energy, force, stress, and property-prediction errors.In short molecular-dynamics trajectories, however, the structural response is strongly system dependent, with some systems showing modest radial-distribution-function changes despite a threefold reduction in inference cost.We then implement and evaluate an owned-core/halo spatial chunking scheme that bounds activation memory during differentiable force evaluation and enables M3GNet simulations of million-atom alloy cells on GPU hardware.These results show that triplet pruning can provide useful acceleration when validated for the target observable, while spatial decomposition offers a more general route to extending GNN potentials to large atomistic systems.
Related
- Experimental Inter-Report Scatter Bounds Machine-Learning Accuracy for TADF Singlet–Triplet Gaps—Structure and Semi-Empirical Features Reach the Same Limit
- Toward Chemical Accuracy in Biomolecular Simulations through Data-Driven Many-Body Potentials: III. A Library of Amino Acid Building Blocks
- Accuracy and Efficiency Benchmarks of Pretrained Machine Learning Potentials for Molecular Simulations
- Diagnosing Latent Energy Decomposition in Machine-Learning Interatomic Potentials via Interacting Quantum Atoms
- Using structure to evaluate the accuracy and transferability of machine learning potentials of biomolecules
BibTeX
@misc{la2026triplet,
title = {{Triplet Pruning and Spatial Decomposition in M3GNet: Accuracy and Memory Trade-offs in Scaling Graph Neural Network Potentials}},
author = {Linh La and Yugam Aggarwal and Truyen Tran and Svetha Venkatesh and Sherif Abdulkader Tawfik},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15001250/v2},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15001250/v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - Triplet Pruning and Spatial Decomposition in M3GNet: Accuracy and Memory Trade-offs in Scaling Graph Neural Network Potentials
AU - Linh La
AU - Yugam Aggarwal
AU - Truyen Tran
AU - Svetha Venkatesh
AU - Sherif Abdulkader Tawfik
PY - 2026
DA - 2026/09/23
PB - ChemRxiv
DO - 10.26434/chemrxiv.15001250/v2
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15001250/v2
AB - Machine-learning interatomic potentials of several functional forms now support large-scale atomistic simulation.Within this landscape, graph neural network (GNN) potentials such as M3GNet offer transferable message-passing models, but their application to very large systems remains limited by graph construction, message passing, and differentiable force-evaluation costs.In M3GNet, the explicit angular pathway is represented through line-graph triplets, whose construction can dominate inference cost.Here we examine the scalability of M3GNet molecular dynamics via inference-time triplet pruning.We first show that an apparent insensitivity to triplet removal in inspected MatGL implementations originates from an incorrect parent-bond index mapping; after correction, removing triplets substantially increases static energy, force, stress, and property-prediction errors.In short molecular-dynamics trajectories, however, the structural response is strongly system dependent, with some systems showing modest radial-distribution-function changes despite a threefold reduction in inference cost.We then implement and evaluate an owned-core/halo spatial chunking scheme that bounds activation memory during differentiable force evaluation and enables M3GNet simulations of million-atom alloy cells on GPU hardware.These results show that triplet pruning can provide useful acceleration when validated for the target observable, while spatial decomposition offers a more general route to extending GNN potentials to large atomistic systems.
ER -