Liberate Trio Characterization Suite

The Cadence^® Liberate™ Trio™ Characterization Suite is the industry’s first unified library characterization system that brings together characterization, variation modeling, and library validation for standard cells, custom cells, multi-bits, and I/Os. The Liberate Trio Suite includes multi-PVT and unified flows that achieve both accuracy and high-speed performance. Its powerful combination of patented technology for generating and optimizing characterization stimulus and parallel processing capability takes advantage of enterprise-wide compute resources, and it leverages cloud resources for an enormous collection of libraries. The Liberate Trio suite is your one-stop-shop for all aspects of standard cell library characterization and validation.

Unified Library Characterization System

The Liberate Trio suite combines our tested and proven characterization suite with some of the most advanced technology available today in a unified library characterization system.

Multi-PVT Flow

The Liberate Trio suite provides a multi-PVT flow to characterize multiple corners in the same run with the resulting libraries maintaining consistency in structure. Vectors and modeling attributes extracted from standard cell circuit analysis are shared among all corners to reduce runtime and ensure the structural symmetry need for static timing analysis (STA) scaling applications. This simplifies the challenge of dealing with an enormous collection of corners across libraries.

Unified Flow

Statistical libraries in Liberty Variation Format (LVF) and nominal libraries can now be generated using a unified characterization run that shares statistical and nominal SPICE process models. This flow eliminates the need to merge libraries at the end of a statistical run and the combined characterization run improves performance.

Machine Learning

The sooner designers have the complete set of corners characterized, the earlier they can start design implementation and close timing. The machine learning algorithms in the Liberate Trio suite enable prediction of critical corners through clustering techniques to determine which corners need to be characterized. The use of machine learning significantly reduces the number of libraries that will need to be fully characterized while ensuring accuracy using ML prediction. With the Liberate ML option’s “prediction out”, intermediate PVTs can be quickly and accurately derived from fully characterized libraries, providing a more comprehensive set of corners to design teams. The Liberate ML option’s “prediction in” also speeds up the entire characterization TAT and reduces the number of CPUs required to deliver a complete accurate characterized library set. The Liberate ML option is shifting left the entire design cycle.

Cloud Enablement

Characterization of large libraries that would normally take weeks can now be turned around in days. Thoroughly distributed and massively parallel, our library characterization portfolio has been fully optimized to run on cloud-based servers by making characterization processes. The Liberate Trio suite can be used on leading cloud service providers or a company’s private cloud and is scalable to over thousands of CPUs.

Advanced Aging

The Cadence advanced aging solution consists of the characterization of stress condition-independent cell libraries that are used by the Tempus aging-aware STA and deliver aging-aware timing and optimization for digital designs. Unlike global derating or corner-based aging characterization, this advanced aging solution, designed from the ground up, relies on the Liberate Trio suite to characterize a generic, stress condition-independent cell library. The Liberate Trio suite’s advanced aging model enables Tempus aging-aware STA to support dependency of timing degradation on aging context, non-inform aging where stress condition is different than operating condition, and arbitrary target usage profiles (segments of supply voltage, temperature, and age) without computing-intensive re-characterization.

Features

• Utilizes a single script to characterize all PVT corners in a library using multi-PVT flow
• Statistical and nominal libraries unified in a single characterization run
• Efficient multiprocessing delivers 3X runtime improvements for library sets
• Machine learning option predicts critical corners and intermediate PVTs and accelerates the overall characterization runtime and hardware requirements
• Advanced-aging characterization of stress condition-independent cell libraries for Tempus aging-aware STA
• Runtime metrics and results monitoring with a sleek new GUI cockpit