Siemens EDA at GOMACTech

Traditional package and PCB analysis tools face significant limitations when applied to advanced 2.5D/3D packaging, including inadequate modeling of fine-pitch interconnects and inability to capture electromagnetic coupling in dense chiplet configurations. The unique challenges of analyzing silicon versus organic architectures—stemming from vastly different material properties, thermal characteristics, and electrical behaviors—require enhanced simulation capabilities beyond conventional methodologies. Additionally, the industry faces significant workflow challenges due to the mixing of silicon-based design and analysis tools with package and system-level tools, creating data translation bottlenecks, model inconsistencies, and design iteration delays that impede efficient PDN analysis and optimization across the complete chiplet-system hierarchy.

This paper presents a comprehensive PDN analysis methodology specifically tailored for advanced packaging technologies: (1) DC analysis for voltage regulation and current distribution optimization with advanced meshing for fine-pitch structures, (2) AC decoupling analysis utilizing 2.5D and 3D electromagnetic solvers for multi-die coupling effects, and (3) transient droop analysis employing time-domain methods optimized for chiplet power delivery scenarios. We demonstrate how traditional Package and PCB analysis tools can be extended for advanced packaging requirements, including specialized modeling for silicon-based architectures and enhanced material property handling for heterogeneous substrates. The approach addresses tool integration challenges by establishing seamless data flow between silicon EDA environments and package-level analysis platforms using Siemens EDA analysis software enabling unified PDN verification across multiple design domains. The methodology's effectiveness is validated through analysis of representative Intel 3DIC designs: a baseline 2.5D design, and a 2.5D design with Embedded Multi-die Interconnect Bridge (EMIB) technology. This work provides a systematic framework for robust PDN design in next-generation advanced packaging solutions while bridging the gap between silicon and package design tools.

Explosive software growth in software-defined and AI-driven electronic systems amplifies non-deterministic failures and late-stage "integration hell." Traditional verification—built for deterministic hardware—falls short in providing timely, objective correctness evidence for adaptive architectures. This presentation introduces a purpose-built verification framework offering continuous coverage from high-level requirements to low-level implementation. It employs a minimalist, repeatable pattern: natural-language requirements become parameterized constraints, hierarchically allocated across black-box, gray-box, and white-box views.

Verification Capture Points (VCPs)—structured artifacts tracking methods, execution history (with failures), evidence, and status—ensure transparent, auditable progress. Interoperable with SysML v2, it uses the standardized model repository as a collaborative "GitHub for systems models," sidestepping MBSE technical debt. Tool-agnostic and federated via open standards, it integrates best-in-class or legacy tools without lock-in, duplication, or migration—enabling early issue detection and reduced integration risk in complex cybertronic systems.

The Siemens Systems software portfolio covers PCB ideation, creation, simulation, analysis, and integration throughout the PCB ecosystem. This session provides an overview of topics which are highly relevant to the A&D industry:

• AMS simulation topics including rad hard AMS simulation
• Worst-case and component stress analyses
• Requirements systems for traceability/provability
• Supply Chain connectivity and early visibility for Engineering

The semiconductor industry is shifting from hardware-defined to software-defined models, increasing demands for chip designers and on the productivity of EDA tools. To address this, Siemens created the Fuse EDA AI System.

Fuse is a secure, purpose-built generative and agentic AI system for semiconductor and PCB design. Centered on openness, it allows customers to integrate proprietary data into custom workflows with flexible on-premises or cloud deployment options. A centralized multimodal data lake drives productivity by leveraging advanced LLMs, RAG, and agentic methodologies, while support for NVIDIA, including NIM and a Nemotron model, further accelerates inference and high-context reasoning.

These powerful AI capabilities are integrated across the entire Siemens EDA portfolio, increasing productivity across the EDA workflow: from HLS to digital verification, to physical verification, and board design. Fuse also accelerates verification and design flows within complex 3D IC environments, significantly boosting user productivity. By embedding intelligent automation throughout the design flow, Siemens is ushering in the next era of EDA.

The semiconductor industry is shifting from hardware-defined to software-defined models, increasing demands for chip designers and on the productivity of EDA tools. To address this, Siemens created the Fuse EDA AI System.

Fuse is a secure, purpose-built generative and agentic AI system for semiconductor and PCB design. Centered on openness, it allows customers to integrate proprietary data into custom workflows with flexible on-premises or cloud deployment options. A centralized multimodal data lake drives productivity by leveraging advanced LLMs, RAG, and agentic methodologies, while support for NVIDIA, including NIM and a Nemotron model, further accelerates inference and high-context reasoning.

These powerful AI capabilities are integrated across the entire Siemens EDA portfolio, increasing productivity across the EDA workflow: from HLS to digital verification, to physical verification, and board design. Fuse also accelerates verification and design flows within complex 3D IC environments, significantly boosting user productivity. By embedding intelligent automation throughout the design flow, Siemens is ushering in the next era of EDA.

The Siemens Systems software portfolio covers PCB ideation, creation, simulation, analysis, and integration throughout the PCB ecosystem. This session provides an overview of topics which are highly relevant to the A&D industry:

• AMS simulation topics including rad hard AMS simulation
• Worst-case and component stress analyses
• Requirements systems for traceability/provability
• Supply Chain connectivity and early visibility for Engineering

Explosive software growth in software-defined and AI-driven electronic systems amplifies non-deterministic failures and late-stage "integration hell." Traditional verification—built for deterministic hardware—falls short in providing timely, objective correctness evidence for adaptive architectures.

This presentation introduces a purpose-built verification framework offering continuous coverage from high-level requirements to low-level implementation. It employs a minimalist, repeatable pattern: natural-language requirements become parameterized constraints, hierarchically allocated across black-box, gray-box, and white-box views.

Verification Capture Points (VCPs)—structured artifacts tracking methods, execution history (with failures), evidence, and status—ensure transparent, auditable progress.

Interoperable with SysML v2, it uses the standardized model repository as a collaborative "GitHub for systems models," sidestepping MBSE technical debt. Tool-agnostic and federated via open standards, it integrates best-in-class or legacy tools without lock-in, duplication, or migration—enabling early issue detection and reduced integration risk in complex cybertronic systems.

Traditional package and PCB analysis tools face significant limitations when applied to advanced 2.5D/3D packaging, including inadequate modeling of fine-pitch interconnects and inability to capture electromagnetic coupling in dense chiplet configurations. The unique challenges of analyzing silicon versus organic architectures—stemming from vastly different material properties, thermal characteristics, and electrical behaviors—require enhanced simulation capabilities beyond conventional methodologies. Additionally, the industry faces significant workflow challenges due to the mixing of silicon-based design and analysis tools with package and system-level tools, creating data translation bottlenecks, model inconsistencies, and design iteration delays that impede efficient PDN analysis and optimization across the complete chiplet-system hierarchy.

This paper presents a comprehensive PDN analysis methodology specifically tailored for advanced packaging technologies: (1) DC analysis for voltage regulation and current distribution optimization with advanced meshing for fine-pitch structures, (2) AC decoupling analysis utilizing 2.5D and 3D electromagnetic solvers for multi-die coupling effects, and (3) transient droop analysis employing time-domain methods optimized for chiplet power delivery scenarios. We demonstrate how traditional Package and PCB analysis tools can be extended for advanced packaging requirements, including specialized modeling for silicon-based architectures and enhanced material property handling for heterogeneous substrates. The approach addresses tool integration challenges by establishing seamless data flow between silicon EDA environments and package-level analysis platforms using Siemens EDA analysis software enabling unified PDN verification across multiple design domains. The methodology's effectiveness is validated through analysis of representative Intel 3DIC designs: a baseline 2.5D design, and a 2.5D design with Embedded Multi-die Interconnect Bridge (EMIB) technology. This work provides a systematic framework for robust PDN design in next-generation advanced packaging solutions while bridging the gap between silicon and package design tools.