Thoughts on IC development, EDA, and hardware design languages.

Take charge of innovation in your IC design flow … build your own tools!

Posted by: Brad Quinton | Posted on: November 17th, 2014 | 8 Comments

Share on LinkedIn60Share on Facebook0Tweet about this on Twitter0Share on Google+0Email this to someone

IC Design is hard.  Really hard. Luckily, IC Design, Verification and CAD Engineers are an extremely talented and creative bunch.  Their job is to innovate everyday.  And they do.  But why should the innovation stop on the device design and verification side?

At Invionics, we believe it is time for IC vendors to take charge of innovation in their IC design and verification flows by building the custom tools they need to get to market faster and with better products than the competition.  We also believe that a platform approach will make this quick and easy to do (but more on that later).

“But wait”, you say, “we have been here before! The major semiconductor players used to build their own tools, but then decided it was more cost effective to outsource the effort to the EDA industry”.  True, but internal development never did go away completely.   Talk to the CAD organizations at any IC company, big or small, and they will quickly name at least a handful of internally developed Perl scripts, spreadsheets, databases, XML parsers, and a whole lot of Tcl configuration scripts that are required to get the design and verification flows they need. Of course, they are not developing their own synthesizers or place and route tools, and they shouldn’t be, but they are developing internal software to get the job done.  It is just part of the job description when you are required to innovate every day in an industry as fast moving as ours.

However, the size, complexity, importance and resource cost of that internal software are all on the rise. Why?  We believe it is the convergence of a number of important trends:

1) The dramatic reduction in EDA startups, especially VC funded EDA startups, has really reduced the variety of EDA tools available today. This is making it more and more difficult to find a way to stand out from the competition with a differentiated design flow.

2) The large and growing cost of IC development means that the cost of internal software is a small part of the overall budget. If you are spending $5M-$10M on a tapeout, it is not hard to spend $100K to develop something that will help make sure you don’t need a re-spin!

3) The continued time to market pressures, especially in the consumer space, justify almost any up-front effort that will pull-in production release of parts. It’s worth it, even if it’s only pulling in by a few weeks — you just can’t afford to miss that slot in the next iPhone.

The trends and needs are real, but it still doesn’t mean that IC vendors can simply hire dozens of seasoned C++ developers to get the job done.  Nor is it reasonable to expect that one more Perl script will lead to innovation.  That is why at Invioincs we have been working hand-in-hand with a number of the largest semiconductor companies in North America to define, develop and deliver a platform specifically for development of custom internal EDA tools, The Invio Platform.  And, today, I’m extremely excited to say that our platform is now available for general use! In fact, you can take it for a spin right here:Free 30-day Invio Trial.

Taking a page from the mobile app development space, the Invio platform is designed to do all the heavy lifting to create robust, high performance EDA tools very quickly. This allows your team to focus on the final bit of “special sauce” (where the true value-add lives) that makes your internal tool different and gives you an edge over the rest of the industry.

The Invio Platform is meant for CAD, Design and Verification Engineers. Invio is equipped with an easy to learn Tcl or Python API, full support (even the non-synthesizable subset) for Verilog, SystemVerilog and VHDL, a Custom GUI Builder, Application Packager, and supports a suite of application specific plug-in modules for RTL Modification, Functional Verification, Netlist Modification and SoC Assembly functionality. Everything you need to get started quickly.

So now the only question left is:  What are you going to build?






Share on LinkedIn60Share on Facebook0Tweet about this on Twitter0Share on Google+0Email this to someone

Comments (8)

  1. Tudor Timi - Reply
    November 19, 2014

    Having dabbled in internal EDA tool development myself, I find your idea awesome. I’ll be following you guys closely!

    • Brad Quinton - Reply
      November 19, 2014

      Hi Timi,
      Thanks for the positive feedback. Now that we have started to look, it is amazing how many people we run into who are also working on internal EDA tools. It’s almost a hidden industry. Please keep in touch. We will be posting lots of new application ideas over the next few months, I’ll be interested in your thoughts.
      Easiest way to keep tabs on us is to follow us on Linkedin.

  2. Naveen Kumar - Reply
    November 28, 2014

    Hi Brad,
    Do you have any user documentation on what issues you are trying to solve with this trial version?

    • Brad Quinton - Reply
      November 28, 2014

      Hi Naveen,
      We don’t have a user document specifically outlining issues that we are trying to solve. We have some ideas on this page: High Level App Examples. But the really cool thing about a platform is that you can really build anything! Essentially all the front-end EDA tools parse and process Verilog, VHDL and SystemVerilog. Everything from synthesizers to simulators, and IDEs to Linters. We also include numerous “lower-level” example applications in our evaluation package: identifying and replacing flip-flops, tracing transitive fanin/fanout, create graphical front-ends, etc.

  3. Tudor Timi - Reply
    December 3, 2014

    Coming back to the idea of apps. Will you also offer a type of marketplace for third party EDA apps that are built using your framework?

    • Brad Quinton - Reply
      December 3, 2014

      Thanks Tudor. The idea is not lost on us! I don’t want to say much more now, but we are very happy to work with any interested company “one at time” for the the moment.

  4. Sriram Kotni - Reply
    December 11, 2014

    I do agree with your thoughts and the direction you want to drive the EDA industry specifically for new innovations in EDA. Today new talent is more interested in social platform development and not in EDA.

    Developing a tool (new idea) from scratch will be costly and time consuming. Having such such platform will be helpful.

    Are you planning to support low power standard IEEE 1801 standard. This will be a useful addition.

    Another good addition, adding to Tudor’s comment on marketplace, is letting individuals and small group of talented people/startups to develop new EDA tools and market them using app store. This will become a gateway to new innovations in EDA industry.

    • Brad Quinton - Reply
      December 11, 2014

      Thanks for your feedback Sriram! And thanks for the suggestion on IEEE 1801. It is certainly on our radar! It would definitely be a useful addition…

      You will have to “stay tuned” on the EDA marketplace front, but I share your feeling that there has to be away for talented developers kick-start new innovation in EDA. I would love to hear from anyone who is in that group. We may be able help!

Leave a Comment