Siemens Accelerates Semiconductor Library Characterization

Semiconductor design teams face increasing pressure to generate accurate timing and power libraries as process complexity, tighter tolerances, and advanced modeling formats expand characterization workloads. In this context, Siemens has introduced Solido™ Characterizer software, extending its Solido™ Characterization Suite with AI-assisted semiconductor design automation capabilities.

Faster library generation for advanced design nodes
Solido™ Characterizer is designed for foundries and internal chip design teams generating SPICE-based Liberty files across both mature and advanced process nodes. The software applies predictive AI methods to accelerate characterization workflows that are increasingly constrained by larger datasets, multiple process-voltage-temperature (PVT) variants, and advanced timing models such as Liberty Variation Format (LVF).

Siemens states that the software reduces Liberty file generation timelines from weeks to days while increasing throughput by seven times. This performance gain comes from a combination of AI-assisted characterization and simulator acceleration rather than from a single optimization layer.

AI and simulation acceleration working together
The performance architecture combines the Characterizer AI Engine with Solido LibSPICE, Siemens’ internally developed characterization simulator.

The Characterizer AI Engine accelerates Liberty generation and LVF characterization, contributing a reported fivefold speed improvement in silicon characterization workflows across process nodes. Solido LibSPICE adds a further performance increase of more than 2x for SPICE simulation workloads used during characterization.

Together, these technologies target a core bottleneck in electronic design automation (EDA): generating validated production libraries quickly enough to match accelerated semiconductor development schedules.

Scaling characterization across design organizations
As semiconductor companies manage multiple IP blocks and parallel design teams, characterization scalability becomes an infrastructure issue as much as a modeling challenge.

The platform integrates with Solido Analytics to provide real-time quality assurance, run monitoring, interactive debugging, and automated replay functions intended to reduce engineering overhead during verification and troubleshooting.

Solido™ Characterizer also connects with Solido Generator, which uses Liberty base files to train AI models capable of generating additional library views without requiring fresh SPICE simulation. Siemens also positions the software for integration with Solido Fuse, based on the Fuse™ EDA AI system, to support generative and agent-based AI workflows across characterization processes.

Customer validation in foundry and IP workflows
GlobalFoundries reported using the Solido Characterization Suite to validate design margins and generate Liberty files while maintaining SPICE-level production accuracy, with internal process acceleration of 20% to 30%.

Anatrix cited the use of Solido Characterizer in characterization of a radiation-hardened digital gate library with single-event latch-up protection, alongside validation of analog and mixed-signal IP behavior through Siemens EDA simulation flows.

For semiconductor teams balancing model accuracy with shorter design cycles, AI-assisted characterization is increasingly becoming part of mainstream chip design infrastructure rather than a specialized optimization layer.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.siemens.com