While the hype around the "Big Data" buzzword has settled down a bit, the amount of data generated during chip design certainly fits the bill. Nowadays, TBs of data get generated during each tapeout, especially at lower process nodes. This presents both challenges and opportunities for chip design teams.
The challenges are well documented:
Data in the Terabytes: The amount of data generated during each ASIC or SoC design is staggering, especially at the latest process nodes. As the feature sizes go down, the amount of data generated per chip goes up, in an exponential manner.
Limitation of On-Premise Centralized Storage: All that data isn't going to fit on your desktop or home directory. The data must be stored in a central location, but finding a centralized data storage solution that can handle hundreds of terabytes of data can be challenging. The most commonly used NFS storage is often a point of contention between design and verification teams, who may not be using the same file system or may be using legacy storage systems that lack the capacity to handle the amount of data being generated. It is not uncommon for designers to send emails to archive or delete past data, a lot of which is extremely useful for future planning and analysis.
Lack of Archiving: How do you organize and archive your data, effectively keeping it organized and available for later use? In our experience, this is one of the biggest challenges facing design teams, and yet the solution is simple: cloud storage.
Data Analysis: How is your team going to make sense of all that data? What's the best way to parse it, store it and analyze it? What tools are best suited for the job?
Fortunately, there are many tools, cloud services, and open source projects available for data storage, data integration and analytics.
While this broad set of options could be overwhelming, we have seen teams gain a competitive advantage by using analytics to learn insights into their design processes, in real time, without the need to go back and re-do past simulations or place and route runs. For example, a team can collect and analyze internal metrics during each cycle of the design process, while they're still working on the project. But this approach only works if you have days or weeks of past design data to compare against. The real value of analytics comes in the ability to spot trends, gain insight and make decisions quickly--when the design is still in flight.
To learn more about how our solution can help you, contact us at: firstname.lastname@example.org.