Product Lifecycle Management offerings have helped companies manage their product lifecycles for decades, reducing costs, improving quality and making customers happy. Surprisingly, though, such a process is not in place for the silicon chips used in almost everything these days. While the semiconductor industry has matured on many fronts, producing vast quantities of faster, cheaper and more power-efficient devices, lifecycle management for silicon has remained mostly lacking. Now, however, EDA vendors are finally building the hardware and software needed to support an end-to-end, data-centric Silicon Lifecycle Management (SLM) solution.
What is SLM?
The SLM process begins with the placement of data collectors and traditional Process-Voltage-Temperature (PVT) sensors on a chip. Analysis of this data can improve yields, quality and enable preventive maintenance in the field. Autonomous Vehicles are the poster child for the need to adopt Silicon Lifecycle Management (SLM). Cloud infrastructure could also greatly benefit from such a platform to improve service quality and reduce costs.
Such a cross-lifecycle solution could provide access to device health data, addressing quality and security challenges through hardware-based electronic maintenance. Some EDA vendors are beginning to offer solutions that designers can use as a start, and more comprehensive platforms remain under development.
The system on a chip (SoC) designer is critical to successful implementations as engineers must embed sensors and monitors into a chip. An SLM platform feeds this data to test systems, engineering teams and customers, building an SLM database. Analytics pulling information from this database can help calibrate the design, identify yield limiters, evaluate and improve product quality and provide predictive maintenance in the field.
Potential benefits of SLM?
Analytics can improve design calibration, accelerate yield improvements, reduce testing time and time to market and, perhaps most importantly, predict failures or quality deterioration in the field. Effects of device aging and new security threats are identified at the hardware level and resolved in the field without replacing the silicon device.
The Synopsys SLM Platform
Synopsys recently introduced a platform for SLM, including technology from recently acquired Moortec (no relation) to speed time to market. This foundation provides the sensors and monitors that feed the database and analytics platforms. Synopsys SLM provides the silicon IP, a lifecycle database and a suite of analytic tools that semiconductor vendors and customers can deploy now to begin the process.
Synopsys is not alone—other SLM platforms are available that provide specific features for specific challenges. One example is ProteanTecs, which focuses on design optimization. Additionally, Mentor/Siemens recently acquired UltraSoC to add embedded monitoring and analysis to its Tessent product suite.
Conclusions and recommendations
To realize the potential benefits, system architects, SoC architects, and their management must embrace the SLM vision for products and define the end-user services. The traditional focus on “Performance, Power, and Area” should be augmented with “Manageability.” With an awareness of the value, benefits and techniques of SLM, IC industry leadership should urge their product design teams to embrace the technology.
The future of SLM is taking shape. Proactive management and design teams are beginning to embrace the concept. While SLM platforms remain a work-in-progress, the time to start is now—the fundamental building blocks are already in place, and several analytic platforms are already available.
I recently published a research paper that gives a more in-depth look at SLM. You can find it here if interested.