NVIDIA cuDSS generally available for GPU sparse-system solves
Original title: NVIDIA cuDSS Now Generally Available to Accelerate Engineering and Scientific Simulations
Vendors report speedups from 4x to 10x in chip, plasma, and structural simulation tools that already use cuDSS.
NVIDIA said cuDSS, a CUDA-X library for large sparse linear systems, is generally available and moves a repeated simulation step onto GPUs. It scales from workstations to clusters and offers C, C++, Python, and Fortran interfaces. Applied Materials, COMSOL with ZEISS SMT, Cadence, Dassault Systèmes, Siemens, and Synopsys say they are integrating it. NVIDIA is showing it at GTC Berlin.
Unconfirmed: Speedups are vendor-reported measurements, not an independent benchmark.
- Applied Materials reports up to 10x faster sparse solves in Ginestra versus CPU-only.
- ZEISS SMT measured a 4x speedup on an EUV plasma benchmark, finishing in six days and 14 hours.
- Dassault Systèmes reports an 8-10x speedup in Abaqus solvers.
- Cadence uses cuDSS in Celsius; Siemens in Simcenter STAR-CCM+; Synopsys in Sentaurus Device.
- Vendors
- NVIDIA
From the original article
2 min read at blogs.nvidia.com
NVIDIA cuDSS, launched today in general availability, is an NVIDIA CUDA-X library that shifts a critical part of engineering and scientific simulations onto GPUs. Complex engineering feats — from advanced chips to aircraft to industrial systems —– depend on simulation to predict how designs will perform before they’re built. As designs become more complex, simulations […]
By Timothy Costa · blogs.nvidia.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.