AI-Generated Constraint Methodology for PCB and IC Package Design Teams

29 Jul 2024 • 5 minute read

It is well-known that constraint-driven designs are correct from the outset and adhere to the physical and electrical rules set during schematic creation. Since the industry's first implementation of constraint-driven design flow in the mid-1990s, the demand for integration between design and analysis tools with automation has grown significantly. Seamless collaboration and data exchange during different design stages is desired and essential for enhancing productivity and efficiency in the modern landscape.

Cadence has launched the industry’s first constraint-integrated flow with AI-based optimization between Allegro and Topology Workbench to improve productivity and enhance the information exchange between different design stages. During the recent CadenceLIVE Silicon Valley 2024, Brad Griffin and Michael Catrambone from Cadence presented AI-generated constraint methodology for PCB and IC package design teams. This blog is an excerpt from their presentation. If you missed the chance to watch this presentation live, please click here to watch it.

Key Challenges in Design and Analysis Tool Integration

The inception of constraint-driven design flows revolutionized PCB and IC package design by introducing a structured approach to managing design rules. However, with the increasing data exchange between different design stages, the demand for integration between design and analysis tools through automation has grown significantly. Today's fast-paced development cycles and the intricacy of modern designs demand an evolution beyond traditional methodologies. Below are the primary challenges associated with the integration between design and analysis tools:

Independent Operation of Tools: Design and analysis tools typically function in silos, leading to inefficiencies.
Divergent Engineering Backgrounds: The users of these tools possess disparate engineering training and technical skills, which can hinder collaboration.
Inefficient Information Exchange: Communication between designers and signal integrity (SI) engineers often rely on outdated methods such as paper trails and spreadsheets.

These challenges delay design cycles and impede the efficiency necessary for complex designs. Addressing these issues requires a paradigm shift toward a more integrated and automated approach.

AI-Generated Constraint Methodology

Cadence has unveiled the industry's first constraint integration flow to address these challenges. This cutting-edge solution harnesses AI-based optimization to connect Allegro PCB Designer, the foremost name in constraint-driven solutions, with Topology Workbench, a robust system-level analysis tool. The solution with Cadence shift-left methodology and AI-driven design Synthesis with constraints automates the constraint communication and makes the whole constraint-driven design process more productive.

Cadence Shift-Left Design Methodology

The ultimate goal of signal integrity (SI) and power integrity (PI) analysis is to improve the design and increase productivity. However, traditional workflows involve a manual, iterative process between design and SI engineers to establish the constraints, which can be time-consuming and prone to errors.

Imagine there is a tool that offers seamless integration, helps generate the SI report early, and can generate the optimized constraints with fewer iterations!

Cadence shift-left design methodology is an effective methodology that shifts the analysis left in the process. It aims to pass SI information back to the layout phase earlier. This approach enhances efficiency by making constraint management more accessible and more streamlined. In Cadence's optimized flow, AI is pivotal in determining the optimal constraints and automating the information transfer for layout and SI analysis. This innovation allows real-time simulation and analysis on the layout canvas, catching potential issues early and improving overall efficiency.

With the optimized flow, the design engineers can modify the constraints and push them back to the layout without manual intervention, even during the design phase, if the simulation results are unacceptable. The main highlights of this solution that make it different from the traditional flow are:

Optimality: The optimized flow includes Cadence Optimality Intelligent System Explorer, a generative-AI-focused multi-physics analysis and technology that helps find the optimized constraint in fewer iterations. It avoids running multiple sweeps and iterations by the SI engineer in determining the optimized constraint. It's like passing the parameters to the AI engine and then asking about the sweet spot (which optimizes the constraint in this case).

Automated constraint communication: The Cadence shift-left design solution uses a simple push button to pass the constraints to the Allegro constraint management instead of emails or other means in traditional solutions. So, the whole process is enhanced, and constraints are loaded inside the layout whenever required to create a new constraint.

In-design analysis: With the in-design analysis, the easy-to-use Sigrity technology is integrated directly into the Allegro canvas (PCB or packaging). This integration enables engineers to screen SI and PI issues, such as IR drop, de-cap inductance, power assessment on the package and identifying the impedance discontinuities on the canvas.

The idea is to find problems earlier so there are fewer problems at the end of the process. Ultimately, the Cadence shift-left solution optimizes the design workflow and empowers engineers to enhance the quality and reliability of their designs with greater ease and precision.

AI-Driven Design Synthesis with Constraints: Constraint Back Annotation

Cadence has enabled AI-driven synthesis with constraints by including Optimality in the optimized flow. Generative AI-driven Optimality helps find the sweet spot in fewer iterations and can do the back annotation with a push button!

The flow offers multiple benefits for the designers, such as:

Offline constraint generation: If the design is not open, an offline file (ACC) can be generated for later usage
Incremental nature: Impacts only the specific extracted nets, ensuring that it does not override all the design constraints and trigger a complete reset
Reusable for design

New Optimized Constraint Generation

This method's technical prowess lies in its ability to use AI for optimally determining the best signal integrity parameters. To achieve the goal, this flow addresses the challenges by following innovation.

Though Allegro and Topology Workbench are different environments, a path is built to link Allegro’s Constraint Manager and simulation results in Topology Workbench.
A generalized GUI form reads in preferred/optimized design parameters generated by simulation, sweeping, or optimization and presents it to layout designers.
The new flow creates design rule sets in constraint manager input format by mapping the constraints manager rules and the simulated parameters from the Topology Workbench.
Users can save the rule set for future designs, update the constraints in an existing design, and check and apply the rule set in Constraint Manager.

The complete process of new constraint creation and back annotation can be done using the following steps.

Create pre-layout topology.
Trace model parameters used for design space exploration.
Setup optimization- Prepare for constraint creation.
Generate optimized constraints.
Check Optimized results.
Prepare Constraints for Allegro constraint manager back annotation.

Conclusion

Cadence AI-generated constraint methodology for PCB and IC package design revolutionizes the design process. It is a unique solution in the industry that embeds the analysis technology directly into the layout environment. This innovative flow allows constraints to be pushed back into the Allegro environment with a push button, allowing engineers to optimize them in real-time and improve productivity and efficiency.