AI BriefingNVIDIA cuDSS Reaches General Availability to Accelerate Sparse Solves in Engineering Simulations
·2026.10.21 19:30
Key point
The CUDA-X library accelerates large sparse systems on GPUs, with partners reporting speedups ranging from 4x to 10x in production workflows.
Details
- NVIDIA cuDSS is now generally available as a CUDA-X library that accelerates large sparse system solves on GPUs, reducing time to results and increasing simulation throughput for engineering and scientific applications.
- Industry partners report significant performance gains: Applied Materials achieved up to 10x acceleration in its Ginestra platform, while Dassault Systèmes reported an 8-10x speedup in Abaqus solvers.
- COMSOL Multiphysics users, specifically ZEISS SMT, measured a 4x speedup in an EUV lithography plasma simulation, completing the full run in six days and 14 hours.
- The library supports C, C++, Python, and Fortran interfaces, allowing developers to integrate GPU acceleration into existing software and scale from individual workstations to large computing clusters.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.