Official source · AI-ASSISTED BRIEF
NVIDIA Daviator library batches k-points to speed up VASP plane-wave DFT
Source publication · YYSI publication
NVIDIA introduced Daviator, a library of GPU kernels that replaces sequential k-point processing in VASP with batched operations, including a batched Davidson eigensolver, targeting small-system plane-wave DFT workloads. NVIDIA reports up to 22x faster time-to-solution and up to 75% lower cost versus CPU execution on a projected OMat24 workload. A mixed precision layer keeps sensitive steps in FP64, claiming roughly 2x end-to-end acceleration on high FP32:FP64 GPUs. Both are available via the ALCHEMI VASP microservice.
Source & context
developer.nvidia.com · Read the original source ↗
This brief uses automated source/date checks and AI-assisted summarization of an official source. Vendor claims are not independent YYSI test results. Check the original announcement before relying on an important detail.