NVIDIA cuDSS now generally available to accelerate engineering and scientific simulations
NVIDIA makes cuDSS, a CUDA-X math library, generally available. It offloads the compute-heavy sparse-system solving step in simulations to GPUs, shortening run times. Applied Materials, COMSOL, Cadence, Dassault Systèmes, Siemens and Synopsys are integrating it; callable from C, C++, Python and Fortran.
What the article says
NVIDIA announced on October 21, 2026, in a post by Timothy Costa, that cuDSS is now generally available. It is a CUDA-X library that moves the solution of large sparse systems onto GPUs, a math-heavy step that often becomes the bottleneck in engineering and scientific simulation. The post starts from the observation that simulations of complex designs take longer and longer, which limits how many ideas teams can test. The stated benefits are faster results, higher throughput and the ability to explore larger, more realistic models.
The post cites vendor examples: Applied Materials up to 10x faster in its Ginestra platform, ZEISS SMT with COMSOL a 4x speedup in an EUV plasma simulation, Dassault Systèmes 8 to 10x for electromagnetic solvers in Abaqus, plus uses at Cadence, Siemens and Synopsys. The library scales from workstations to large clusters and offers C, C++, Python and Fortran interfaces, so software vendors can integrate it into their tools. The figures are cited examples, with no overall benchmark.
Summary written by AnnuIA in its own words from the official page of NVIDIA. The content and rights belong to the publisher: please read the original article.
More announcements from NVIDIA
- Agentic AI joins the search for extraterrestrial signals
- Thin, light and built for everything: NVIDIA RTX Spark Windows laptops gain new models, agent tools and games
- How AI is changing the pace of discovery for millions living with dementia
- Leading European financial services organizations embrace open models to build proprietary intelligence