I think everyone in the EDA industry understands that real innovation rarely goes on in the large established companies – we look towards the startup companies for significant changes in direction or new ways to approach existing problems.
We also hear a lot about how startups are having a tough time at the moment because the number of them being acquired has gone down and there is little venture capital money coming into the industry. I have also noted a number of startups that have almost entered zombie stage – not able to get past a significant sales rate, not showing enough growth to be of much interest, but then again earning enough to just keep them ticking along.
I was pleased to see a number of new companies at DAC this year as this was a sign that startups were beginning to see new opportunities for innovation and one of my favorite areas recently has been FPGA prototyping. Thus I became interested in Veridae Systems
So let’s start with a little bit of background about them. This is a company that was formed in 2009 based on research from the University of British Columbia. Now there have been some great companies that have come directly from university work, but many of them are heavy on the technology and way too light on the business acumen to succeed. So what about Veridae? They were started using private money and never received any venture capital. They brought in management from companies such as PMC-Sierra and SiGe Semiconductor. Their first product was announced at the end of 2010. In June 2011 they announced an extension of the products to cover additional use cases, just in time for DAC.
Veridae is one of the companies taking a software approach to gain greater visibility into what is happening inside an FPGA. By software approach I mean that they create a small “debug engine” that goes into the FPGA – so the solution uses hardware, but not a dedicated hardware board that attaches to your prototype. That engine is responsible for capturing information, compressing it and getting it out in an appropriate manner to the outside world and reconfiguring the debug state of the machine without requiring a lengthy recompilation time.
Now there were quite a few companies claiming this kind of capabilities at DAC and I was eager to learn more about what made their approach different from the others. One of the interesting problems that has to be overcome is that it is no longer sufficient to just get data out of a single FPGA. Many prototypes today contain multiple FPGAs and debug has to work seamlessly across all of them. That requires breakpoint synchronization amongst other control issues and trace synchronization, especially difficult when multiple clocks are in use.
Before I could really get into the company and their products, I saw last week (July 5th to be exact) that they are being acquired by Tektronics
(Tek), a maker of test equipment. They are slated to become the Embedded Instrumentation Group within Tek. This is an interesting move for them. Traditionally Tek has produced equipment that tests a completed product, but this becomes a move into testing the functionality of a system and moreover before the design has been completed. So this acquisition takes Tek in two new and different directions – functional verification and moving forward in the design flow. Now part of the reason for doing this could be that both Xilinx and Altera provide capabilities for the debug of a single FPGA. Agilent, one of Tek competitors, formed a close alliance with Xilinx for test solutions that may have made it difficult for Tek to make headway in this area. This acquisition may give them more clout with the FPGA companies and over time may give them an entry into the ASIC test market, especially for those companies that want to migrate designs from FPGA to ASIC.
It will also provide Veridae with some of the money and marketing clout they need to survive in this market. Several companies have been before them: FS2 a provider of embedded test and debug IP (who also worked closely with Tek) was acquired by MIPS and have gone very quiet since then, and DAFCA, a post silicon validation and debug company appears to have turned off the lights. So while the need and the value of this technology seems self evident, there is clearly a reluctance of designers to allocate even a few percent of their area to providing better visibility, and basically taking out an insurance policy on their design. By using the FPGA prototyping market as a launch pad, perhaps Veridae and other will be able to gain more traction and prove their worth before trying the harder sell of getting the ASIC market to embed this technology into their chip, where the cost is then born in each and every product. This will be a very interesting field to watch.
Brian Bailey (http://brianbailey.us
) – keeping you covered
If you liked this article, go to the Test & Measurement Designline home page for the latest on instrumentation design, technology, trends, products and news.