Monday, September 15, 2008

Alan Pita, President of Logic Refinery, Inc. Talks about System Verification

What is the problem Logic Refinery is trying to solve?

More than 50% of system-on-a-chip (SoC) designs come back from their
first tape-out event with hardware that simply doesn't work well enough to ship. The multimillion-dollar engineering costs associated with doing another tape-out (called a "re-spin") are typically dwarfed by the cost of lost time-to-market.

How are you solving that?

The common approach to verification is to write a software program called a "testbench" to verify virtual models of the chip using software simulation, and sometimes hardware emulation. This approach worked very well for simpler designs. The industry has discovered that it simply doesn't give verification teams enough leverage to verify the more complex SoC designs in a reasonable time.

Our product is a software program we have named the "Strategen Verification Engine". It's the first tool designed to fully solve the system verification problem, and it's vastly more effective than the testbench-only approach. Applying Strategen to their verification plan accelerates and enables the whole system verification effort, allowing all of the critical bugs to be found and fixed before the
first tape-out.

How will your customers estimate the value of our product?

Applying Strategen does have a substantial productivity benefit; however, the real value of Strategen is not merely a function of team acceleration. For the management of SoC design firms, verification is seen as a risk management game. The industry is at a crisis point because they simply can't make the risk of failed SoC tape-outs go away with staffing, schedule, and existing tools. The approach behind Strategen has the ability to close this business risk for management, enabling them to plan for "first silicon success".

Who are you targeting to buy the system?

Firms that are doing system-on-a-chip design, especially the more complex SoC designs with hundreds of IP blocks. There are hundreds of these firms, with thousands of tape-out events every year.

Do you have any customers so far?

We are presently scheduling demos and initiating Alpha testing at customer sites. We'd like to convert some of these Alpha sites into paying customers over the next 6 months or so.

How much will you need to get the company up and running?

Our advisors estimate between four and eight million dollars will be needed in several funding rounds to take us all the way to exit. We’ll start with a first round of about $1.5M USD to get the product and associated business processes prepared for a full-on sales and marketing effort and get the first few paying customers signed up.

What is your beachhead market –say the first 20 customers you’ll go

Basically there are 195 ARM licensees. These are our ideal customers. Again, the verification problem is something that all of our customers recognize and they can't spend enough money to solve it.

Best regards,
Hall T.

No comments: