Mastering AI Chip Complexity
Explore challenges and solutions in AI chip development
Download eBook
Definition
SmartScaling is a methodology that uses advanced machine-learning algorithms to reduce the actual simulation runtime, thus mitigating the above challenges in characterizing multiple PVT corners.
With advancements in technology and the push towards smaller nodes from 7nm down to 3nm, designs are expected to work under different modes, with possibly different clock frequencies along with a range of global variations. Thus, designs have to achieve signoff closure for timing and power across an enormous number of process, voltage, and temperature (PVT) corners.
Characterizing this huge number of PVT corners, library teams across the industry face further challenges like high simulation turnaround time, database disk space limits, license server overloads, and hardware costs.
How Does SmartScaling Work?
SmartScaling provides ML-based adaptive library characterization and scaling of libraries for multiple PVTs. It is a new methodology that is built on top of the PrimeTime® scaling engine, which uses existing characterized libraries as anchor PVT corners to then generate multiple additional PVTs instantly.
SmartScaling provides signoff accuracy for various library views – NLDM, NLPM, CCST, CCSN, and LVF.
Two primary methods for SmartScaling are shown below.
Method 1
- SmartScaling-based PVTs generated via an existing characterized database, used as anchor PVTs.
- Here, based on the anchor corners provided by the customer, the SmartScaling engine will generate additional libraries at multiple PVT corners.
Method 2
- SmartScaling-based characterization for overall multiple PVT corners, where the engine will decide which PVT corners to fully characterize and do additional limited sparse corner characterization. This approach essentially creates a database from which to generate additional PVT corners instantly.
- 20 PVTs = 5 Anchor Char + Sparse Char + 15 SmartScaling PVT corners
The Benefits of SmartScaling
SmartScaling helps to reduce the overall runtime for generating multiple PVT corners for library characterization. Users can:
- Reduce library characterization runtime by 3x to 10x
- Reduce library size by 3x to 10x
- Ensure library accuracy
- Enjoy zero cost / instant additional PVT corners library generation
SmartScaling and Synopsys
Synopsys offers SmartScaling as part of its advanced library characterization solution to address the growing complexity and demand for faster, more efficient PVT corner generation. SmartScaling is integrated with PrimeLib, a unified library characterization and validation platform.
With SmartScaling, PrimeLib enables customers to reduce library characterization runtime and storage requirements by leveraging machine learning to scale from anchor PVTs. This approach provides signoff-accurate libraries for use with PrimeTime static timing analysis, supporting advanced modeling formats such as NLDM, NLPM, CCSN, CCST, and LVF.
SmartScaling further enhances PrimeLib’s capabilities by enabling instant generation of additional PVT corners with zero additional simulation cost, while maintaining high accuracy and reducing reliance on compute-intensive characterization.
SmartScaling supports:
- Cloud-ready deployment, allowing high scalability and flexibility across cloud and on-premise environments.
- Accelerated throughput for advanced nodes (e.g., 7nm to 3nm) by combining SmartScaling with PrimeLib’s optimized flow.
- Compatibility with PrimeTime POCV, AWP, and EM analysis, ensuring accurate and efficient signoff.
Synopsys’ SmartScaling technology empowers library teams to meet time-to-market demands efficiently, without compromising quality, and is an essential tool for modern, ML-enhanced library characterization strategies.
Continue Reading
Blog
5 min read
/
Aug 14, 2025
How AI is Revolutionizing Analog and Digital Node Migrations
Blog
3 min read
/
Aug 11, 2025
AI Vision: Creating Chips Today for Tomorrow’s Cars and Robots
