NVIDIA Daviator library boosts VASP plane‑wave DFT up to 22× on GPUs
New Daviator kernels and mixed‑precision mode cut VASP time‑to‑solution dramatically, lowering cost versus CPU runs.
NVIDIA announced Daviator, a library of GPU kernels that replaces VASP's sequential k‑point processing with batched operations.
The batched Davidson eigensolver reduces kernel launches and improves SM occupancy, addressing the small‑system GPU bottleneck in plane‑wave DFT.
Benchmarks on a projected OMat24 dataset show Daviator‑enabled VASP can be up to 22 times faster than CPU execution, with cost reductions of up to 75 percent, depending on hardware and configuration.
Ce que cela signifie pour vous
Researchers generating massive DFT datasets for machine‑learning interatomic potentials can now run smaller unit‑cell calculations faster and cheaper on modern GPUs. For users of free or locally hosted AI models, the reduced cost of generating high‑quality training data may lower barriers to developing open‑source models.
Sources
- Tackling GPU Bottlenecks in Plane-Wave DFT with K-Point Batching and Mixed Precision, NVIDIA Developer (official), 18 Nov 2026
Rédigé automatiquement à partir des sources ci‑dessus, puis vérifié par notre code : chaque nombre et chaque nom doivent y figurer, et aucune phrase ne doit copier une source. Une erreur ? Signalez‑la.
Quelle IA pouvez‑vous utiliser gratuitement aujourd’hui ?
Tous les modèles IA gratuits, classés par votes humains anonymes, avec la mémoire et la vitesse dont ils ont besoin sur votre propre GPU.