By Michael White
When you’re flying, it’s fun to look out the window and see clouds from “the other side.” When you’re running an IC design company, clouds take on a whole new perspective as well. Just like planes made it possible to cross continents and oceans in hours rather than days, EDA in the cloud can make it possible to get to tapeouts faster. But, just as we trust planes to carry us safely to our destinations, we also have to trust that running EDA software in the cloud will give us the same quality of results we depend on when using on-site resources.
So, why consider Calibre in the cloud? That’s probably the easiest question anyone could ask me! How many reasons do you want?
- Check-driven computation time and complexity is growing node over node, extending runtimes and requiring more compute resources
- Physical verification functionalities continue to expand, requiring additional time and compute resources
- Limited on-site hardware resources may result in long runtimes
- Access to on-site hardware resources is often strictly controlled or allocated, leading to long wait times
- On-site hardware acquisition, installation, and maintenance take time, money, and personnel
But, you might say, what about security? And quality?
- Cloud providers have implemented strict security controls to protect IP data
- Foundries have accepted cloud computing
- Calibre in the cloud uses the same Calibre engines you depend on today, providing the same quality of results you’ve trusted for years
- You can use the same licensing structure you already have in place
Shortening the time to market has always been a critical competitive differentiator for design companies, but according to surveys in the last few years, as many as 50% of tapeouts have missed their scheduled deliveries! Running highly scalable Calibre tools in the cloud can enable companies to perform multiple design iterations per day, or supply an “infinite” resource pool during crucial surge points in the design and verification flow. Just think of the value of moving from two design iterations per day to four—how much faster could you get to tapeout?
And it’s not a theoretical question. We’ve been using Calibre tools in the cloud internally for over five years, accessing hardware pools up to 10k cores. We’ve developed and tested best practices for cloud execution, and fostered collaborations with strategic cloud partners such as AMD, Microsoft Azure, and TSMC. Our relentless focus on performance and memory enhancement is just as important in cloud usage as it is for on-premise resources.
If you’re interested in learning more about the kind of performance and results you can expect from Calibre and cloud computing, here are some references to get started:
If you have specific questions, or would like to get started, contact a Calibre representative to get all the details!