Making Analytic Double-Hybrid DFT Gradients Practical: A High-Performance Multi-GPU RI Implementation

Abstract

Double-hybrid density functional theory (DHDFT) is the most accurate rung within the density-functional hierarchy, but its use for molecular structure and dynamics remains limited by the high computational cost of analytic nuclear gradients. Here, we present, the first GPU-accelerated implementation of analytic DHDFT gradients. The implementation, developed in the Extreme-scale Electronic Structure System (EXESS), employs resolution-of-the-identity (RI) techniques throughout the calculation and evaluates the complete relaxed gradient on GPUs, including the modified Z-vector response and the explicit nuclear derivative of the exchange–correlation potential. The distributed multi-GPU formulation is performanceportable across NVIDIA and AMD architectures, supports basis functions through g angular momentum, and enables calculations through quadruple-ζ basis quality. The implementation is validated against numerical derivatives and the established CPU implementation in ORCA for B2PLYP and revDSD-PBEP86D4. For sufficiently large calculations, EXESS sustains more than 50% of the theoretical peak double-precision floating-point performance and exhibits efficient strong scaling across multiple NVIDIA H200 GPUs and AMD MI250X GCDs. Compared with ORCA on a modern Sapphire Rapids CPU node, a single H200 GPU provides speedups ranging from approximately 10× to 675×. These results extend high-performance DHDFT from single-point energetics to efficient exploration of molecular potential-energy surfaces at the large basis-set quality required for accurate perturbative correlation.

Analytic Nuclear Gradients and Derivative Couplings for SA-QED-CASSCF
Nam Vu, Jonathan J. Foley
ChemRxiv5 Oct 2026PreprintCC BYFull textAtomic, molecular & optical physicsChemRxiv
Implementation and preliminary evaluation of a practical digital workflow for routine laser-radiation field coincidence quality assurance on the CyberKnife M6.
Linzheng Hu, Yongkang Zhao, Chunling Zheng +6
Precision radiation oncology24 Oct 2026Peer-reviewedCC BYFull textApplied physics & instrumentationEurope PMC
Making the Discrete Continuous: Synthetic RAW Augmentations for Fine-Grained Evaluation of Person Detection Performance in Low Light
Valeria Pais, Malena Mendilaharzu, Daniele Faccio +3
arXiv31 Aug 2026PreprintCC BY-SA 4.0Full textOptics & photonicsarXiv
Performance of the spin qubit shuttling architecture for a surface code implementation
Berat Yenilen, Arnau Sala, Hendrik Bluhm +2
Quantum30 Sep 2026Peer-reviewedCC BY 4.0Full textQuantum physics & quantum informationCrossref
Phase diagram morphology shapes droplet propulsion in chemical gradients
Stefan Köstler, Malcolm Steen, David Zwicker
arXiv21 Sep 2026PreprintarXiv licenceFull textBiological & medical physics Condensed matterarXiv

Cite

BibTeX RIS

BibTeX

@misc{pure2026making,
  title = {{Making Analytic Double-Hybrid DFT Gradients Practical: A High-Performance Multi-GPU RI Implementation}},
  author = {Ross Pure and Giuseppe M. J. Barca},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15008642/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008642/v1},
  note = {CC BY-NC-ND},
}

RIS

TY  - UNPB
TI  - Making Analytic Double-Hybrid DFT Gradients Practical: A High-Performance Multi-GPU RI Implementation
AU  - Ross Pure
AU  - Giuseppe M. J. Barca
PY  - 2026
DA  - 2026/09/10
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15008642/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008642/v1
AB  - Double-hybrid density functional theory (DHDFT) is the most accurate rung within the density-functional hierarchy, but its use for molecular structure and dynamics remains limited by the high computational cost of analytic nuclear gradients. Here, we present, the first GPU-accelerated implementation of analytic DHDFT gradients. The implementation, developed in the Extreme-scale Electronic Structure System (EXESS), employs resolution-of-the-identity (RI) techniques throughout the calculation and evaluates the complete relaxed gradient on GPUs, including the modified Z-vector response and the explicit nuclear derivative of the exchange–correlation potential. The distributed multi-GPU formulation is performanceportable across NVIDIA and AMD architectures, supports basis functions through g angular momentum, and enables calculations through quadruple-ζ basis quality. The implementation is validated against numerical derivatives and the established CPU implementation in ORCA for B2PLYP and revDSD-PBEP86D4. For sufficiently large calculations, EXESS sustains more than 50% of the theoretical peak double-precision floating-point performance and exhibits efficient strong scaling across multiple NVIDIA H200 GPUs and AMD MI250X GCDs. Compared with ORCA on a modern Sapphire Rapids CPU node, a single H200 GPU provides speedups ranging from approximately 10× to 675×. These results extend high-performance DHDFT from single-point energetics to efficient exploration of molecular potential-energy surfaces at the large basis-set quality required for accurate perturbative correlation.
ER  -

Abstract

Related

Cite