Will IP use increase in forthcoming SoC design? - Embedded.com

Will IP use increase in forthcoming SoC design?

To answer the question, you have to start with what is driving the future of system on chip (SoC) design. Unless we all suddenly decide to stop buying the next cool electronic gadget on the market, and the one after that, consumer electronics will continue to be the biggest consumer of SoCs. The electronic content in our daily life keeps growing, and so does the need to stay connected wherever we happen to be on the planet. This will continue to fuel the growth of consumer electronics and SoC design.

The battleground in consumer electronics continues to be time-to-market, software applications, user experience and cost, not always in that order depending on the end-market. You can lose upwards of 50 percent market share if you are late to the market by a few weeks, and in some cases be completely shut out – Christmas does only come once a year after all! As a result, SoC design schedules usually are extremely aggressive and deadlines often non-negotiable. On the other hand, the software content in consumer electronics devices is exploding and so is the required computer horsepower. What used to be a single processor now has been replaced with dual or quad processors, graphics processors, microcontrollers, DSP, etc. A typical Smartphone today has more computing horsepower than a mid-size desktop just a few years ago. Memories on-chip and off-chip (and we are talking hundreds of Gigabytes now) are another major requirement – not just the amount of storage, but rapid access to it as well. Standard interfaces are yet another aspect of any consumer electronic chip where the device must plug into various standards – USB, SATA, HDMI, PCI-X, and so on. And to make sure all of these components are able to communicate with each other you need bus fabrics to handle all the traffic.

So if you are an SoC designer, you have the unenviable job of bringing all of this together, and on top of that, having to focus on differentiation over the next guy building a similar chip and trying to beat you to the market. But you cannot possibly differentiate on every functional block and still meet your deadlines. The use of semiconductor IP is a viable solution to this seemingly insurmountable task. Whether developed internally by a central group or acquired from a commercial vendor, the use of IP allows you to focus on things you differentiate on and to farm out all other general purpose blocks. Studies have shown that an effective IP strategy can cut down the overall SoC design effort by as much as 50-60 percent.

So why do SoC designers have such a nightmare using IPs delivered by 3rd parties or internal IP teams? A closer look highlights a few issues:

1.   The barrier to entry into IP market is too low – Too many companies have jumped into the IP business with little or no expertise in the specific domain. Too many IP suppliers compete for the same SoC business thus diluting market-share and limiting the ability to invest in core competence, tools and processes. The industry needs more specialized IP vendors that focus on a specific class of IP instead of a set of me-too solutions.

2.   Lack of enforceable quality metrics and tools – IP vendors focus on verifying functionality but put little to no focus on ensuring the implementation readiness of the IP. There is definitely a need for well-defined quality metrics and tools to address this problem.

3.   IP integration challenges – Most IPs are parameterized and can be configured for the specific needs of an SoC. IP vendors typically validate the IP for a fixed few configurations, but there is no guarantee it will work in the SoC context. This can be bridged with a combination of tools, services and a business model whereby the IP vendor shares the risks and rewards of making their product work in the SoC context.

4.   Drive-by IP reuse – Many SoC teams try to reuse design blocks from the previous generation’s design. If the original block was never designed for reuse, integration is going to be challenging. There needs to be some effort put into characterizing and packaging legacy IP blocks before they are re-used in an SoC.

The debate going forward is not so much whether IP use will grow, but how effective will it be? In other words, what is the return on investment? For the consumer electronics market, the payoff can definitely be huge. The companies that have already put together a complete IP strategy are reaping the benefits and will continue to do so. Others might find themselves playing catch-up.

Speaking as a consumer, I will keep an eye out for the next cool gadget on the market!

About the author:
Piyush Sancheti is currently the senior director of business development at Atrenta.

Piyush is responsible for Atrenta's strategic alliances with key members of the semiconductor supply chain. He has been in the EDA industry in various marketing and product management roles since 1993. Prior to Atrenta, Piyush held marketing and engineering positions at Sequence Design, Senté, and Cadence Design Systems. Piyush holds a Masters in Computer Engineering from Iowa State University. He can be contacted at .

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.