• Skip to main content
  • Skip to search
  • Skip to footer
Cadence Home
  • This search text may be transcribed, used, stored, or accessed by our third-party service providers per our Cookie Policy and Privacy Policy.

  1. Blogs
  2. Verification
  3. Xcelium ML: The Next Big Thing in Regression
XTeam
XTeam

Community Member

Blog Activity
Options
  • Subscribe by email
  • More
  • Cancel
machine learning
xcelium
Regression

Xcelium ML: The Next Big Thing in Regression

1 Sep 2020 • 1 minute read

Looking for that extra kick in your regression performance? Cadence’s Xcelium Logic SimulatorTm has a new feature just for you. Harnessing the power of machine learning, which is one of the areas of computational software innovation, Xcelium ML is here to help you optimize your regressions. 

The inherently iterative, data-driven nature of simulation seems ripe for a machine-learning assisted tool, and Xcelium ML is here to fill that gap. Xcelium ML’s goal is to create a positive feedback loop in the simulation progress, ensuring that there’s no dead time on the part of the simulator doing the heavy lifting or the engineer creating the tests. 

Xcelium ML is an interface that attaches to your existing Xcelium installation. The Xcelium ML gathers data about coverage and the random seeds used in the user’s regressions as they’re performed. It then analyzes the influence of these conditions and others as a factor of how well these runs achieve the target coverage goal, creating a superior set of instructions as a part of a model. These conclusions are refined across multiple regressions, matched up to statistical models, and are eventually used to generate a leaner regression that hits the same coverage, but much faster. All of this is done transparently, so a user can see what data Xcelium ML is gathering as it collects it, and it’ll provide the user with information about which tests are used in a given analysis and how that affected the regression. This is all customizable, too, so an engineer can modify the target coverage, the ratio of directed tests versus random tests, and so on. 

Now, let’s look at an example. 

In a customer’s regression—17,000 runs, 235 tests, 32,000 bins hit in about 4,000 CPU hours using regular, unassisted regressions, Xcelium ML allowed them to hit their target coverage of 99% in just 5,800 runs and 1,000 CPU hours, using all random seeds for both. To hit that last 0.9%, another regression was performed, this time with three times as many runs—and Xcelium ML beat the normal regression’s time by 33%. And, of course, all of this is done while providing the user with comprehensive, diverse analytics to let an engineer in on exactly what’s going on at every step of the process, easing debug. 

So what are you waiting for? Try Xcelium ML today, and don’t waste any more time doing regular regressions. Experience what Cadence’s state-of-the-art computational software can do for your design cycle. For more information, check out the Xcelium page. 

© 2025 Cadence Design Systems, Inc. All Rights Reserved.

  • Terms of Use
  • Privacy
  • Cookie Policy
  • US Trademarks
  • Do Not Sell or Share My Personal Information