GS-DFT scales DFT with Gaussian Splatting, hits 2,742 atoms on four H200 GPUs
AllThingsApx · x · 2026-09-28
A new paper (arXiv:2609.31483) introduces GS-DFT, representing molecular orbitals as a cloud of Gaussians whose positions, shapes, and mixing coefficients are jointly optimized by gradient descent to minimize energy — no training data needed. Conceptually it's "3D Gaussian splatting with the renderer replaced by quantum mechanics."
- Key components: adaptive density fitting with screening for two-electron integrals, and regularized differentiable orbital orthogonalization
- Accuracy: matches the largest conventional basis sets with a fraction of the parameters, converging systematically in energy, density, and forces; captures stretched-bond and anion physics at equal parameter count
- Scale: quadratic peak memory scaling allows simulating up to 2,742 atoms (10,406 electrons) at triple-zeta on a single 4×H200 node
Paper and code are public; the next frontier is reliable forces at lower compute cost.
Related event: GS-DFT Scales Quantum Chemistry to 2,742 Atoms via Gaussian Splatting(4 posts)→
More from Research
- Continual-ARC: lifelong ARC-AGI stream finds agent harness persistence rarely beats simple caching — unifyai · 2026-11-16
- DySENSE: MIT's GenAI framework synthesizes satellite imagery with dynamic city energy maps — MIT · 2026-11-16
- WebArena-Pro: a 300-task multimodal benchmark across 20 self-hosted apps for web agents — McGill-NLP · 2026-11-16
- Training-free page-sparse decode in llama.cpp: 1.72× faster 128k decoding on an RTX 5090 — topppggg · 2026-11-16
- Training Qwen3-8B to drive a simulated hand spawns a dedicated internal 'pressure' direction — repligate · 2026-11-16
- Google Research open-sources PaperOrchestra, a multi-agent framework for writing AI papers — google-research · 2026-11-16