2024 predictions for EDA at Keysight
EDA is also turning to AI says Niels Faché, Vice President & General Manager, Design and Simulation at Keysight
“The application of AI and ML techniques in EDA is still in the early adopter phase, with design engineers exploring use cases to simplify complex problems. The intelligence is particularly valuable in model development and validation for simulation, where it assists in processing large volumes of data,” he said.
“In 2024, organizations will increasingly adopt both technologies for device modeling of silicon and III-V semiconductor process technologies, as well as system modeling for forthcoming standards such as 6G, where research is well underway.”
New standards such as UCIe are emerging for the creation of chiplets and the disaggregation of system-on-chip designs into smaller pieces of intellectual property that can be assembled into 2.5D and 3D integrated circuits using advanced packaging. For designers to accurately simulate die-to-die physical layer interconnect, it will require high-speed, high-frequency channel simulation to UCIe and other standards.
Connecting the digital and physical worlds will require more powerful digital processing and interfaces able to overcome increasingly complex signal physics. An array of advancements in semiconductor technologies will be essential to achieve this and overcome the associated challenges.
“These issues include increasing data rates that need wider bandwidths, which dictate higher carrier frequencies, extending into the THz regime for wireless. The use of techniques such as extreme MIMO, adds more complexity and density, and networks with diverse topologies, such as the use of non-terrestrial (satellite) links, magnifies the challenge,” said Dan Thomasson, Head of Central Technology and Vice President of Keysight Labs.
“Innovations to address this will include combining commercial semiconductors, such as GPUs and FPGAs, with custom MMICs and ASICs, and these new solutions will deliver significant improvements in size, weight, performance, and power consumption. Data converters enabling the capture and generation of signals at the widest bandwidths with unsurpassed signal fidelity will be needed. In addition, photonic solutions will be critical to extend the reach and capacity of data transmission technologies.
Workflows also need to be addressed. Workflows are currently a set of loosely connected tools. However, as the virtual and physical worlds merge, a unified design and test workflow where data is shared seamlessly via the cloud between simulation and measurement steps is required. The information will be constantly analyzed to inform the behavior of simulation and measurement, eliminating any gaps in the workflow from concept to final test. The insights from the simulation will be fed into AI-driven tools that will elevate the speed and productivity of the design and test workflow. Digital twins will be used to tightly couple design and test so that only one actual build is needed.
As design moves from the physical into the virtual space, engineers are able to discover and fix problems in the most efficient manner, providing greater insights and performance improvements, says Fache. The next few years will see a continuing emphasis on connecting design and test workflows to handle rising complexity and more demanding time-to-market requirements for electronic products in wireless, wired, aerospace & defense, and other industries.
Software automation techniques, such as Python APIs, will take a more significant role in integrating “best-in-class” tools into open, interoperable design and test ecosystems.
With the creation of digital enterprise workflows, many organizations are investing in design management across tool sets, data, and IP. Moving forward, design data and IP management software will play a critical role in the success of complex SoC and heterogeneous chiplet designs supporting large, geographically distributed teams. The creation of digital threads between requirements definition and compliance as well as establishing tighter links with enterprise systems such as PLM will play a role in the digital transformation of product development cycles.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
