Siemens brings Nvidia GPU acceleration to Calibre lithography simulation

Siemens is integrating its Calibre manufacturing software with Nvidia's cuLitho library, which runs computational lithography on GPUs. The companies say full chip optical proximity correction, the step that adjusts mask shapes so printed features come out as designed, runs up to 12 times faster, and Siemens reports verified speed ups of between 10 and 17 times across the main lithography simulation workloads. The work also covers curvilinear inverse lithography technology, which uses freeform mask shapes and becomes more practical for High NA EUV when run times fall. Faster simulation lets chipmakers and mask shops run more optimisation passes, find manufacturing problems earlier and improve mask accuracy before tape out, which matters most for large AI accelerators and advanced memory where masks are expensive and process windows are narrow. The announcement does not give an availability date, licensing model or customer names, so it is not yet clear when production users will be able to run the accelerated flow.

Source: Siemens and NVIDIA speed up AI chip lithography by 12x, eeNews Europe, published 2026-10-16. The summary is the Hub's own; the full story is at the source. Report a rights or accuracy concern