Solido ML Labs

Machine Learning-Based EDA Software Solutions

 

Now Available to Semiconductor Companies!

We’re bringing our machine learning for engineering technologies, our expertise, and our customer support structure together to design and develop innovative, new customer-driven EDA design solutions!

Solido has the machine learning for engineering technologies and expertise to address:

  • Expensive time- and resource-intensive data acquisition problems generating overwhelming amounts of data
  • Big data problems requiring expert resources for data insights and interpretation
  • Data visualization problems requiring advanced user interfaces
  • Complete EDA product solutions to integrate in current design flows

We’re looking forward to collaborating with you! Our experts will be with you for every step along the way: identifying challenges, innovating solutions, developing tools, and implementing the final product, so it ‘just works’ in your design flow.

Contact us at mllabs@solidodesign.com and we’ll be happy to discuss solutions that will work for you.

ML Labs Product Successes

Launched in 2017

Predictor

Predictor

  • Uses machine learning to accelerate characterization of standard cell, memory, and I/O
  • Reduces characterization runtimes by 30% to 70% without compromising accuracy
  • Saves days to weeks in characterization time
  • Saves on SPICE simulator and characterization resources. 

Launched in 2017

Statistical Characterizer

Statistical Characterizer

  • Uses machine learning technology to accelerate statistical characterization of standard cells, memory, and I/O
  • Generates statistical library models with Monte Carlo and SPICE accuracy +1000x faster than brute force Monte Carlo
  • Precisely handles non-Gaussian distributions
  • Delivers true 3-sigma LVF/AOCV/POCV

ML Labs Core Technologies Based on the Following Products

Variation Designer 4

Variation Designer 4

The world’s best technology for variation-aware design of memory, analog/RF, and standard cells

ML Characterization Suite

ML Characterization Suite

Fast, accurate library characterization tools powered by machine learning

Features of our machine learning technologies include:

  • Production-accurate results, not estimations
  • Self-verifying adaptive model building to correct errors
  • Verifiable and scaleable to 100K+ input variables
  • Parallelizable to large clusters
  • Captures high-order interactions, non-linearities, and discontinuities
  • Used by top semiconductor companies