Blogs

We wanted to thank another one of our IMPACT sponsors - OCP-IP.

"OCP-IP is dedicated to proliferating a common standard for intellectual property (IP) core interfaces, or sockets, that facilitate "plug and play" System-on-Chip (SoC) design."  Check out their website by clicking on their logo.

Grant Pierce, Chief Executive Officer, Sonics, Inc., gave a keynote presentation at the Semico Impact Conference: Focus on the IP Ecosystem on May 16, 2012.  His presentation that looked at the rise of cloud computing and its impact on the SoC market. Today, consumers expect connectivity ‘Anytime, Anywhere’ and can mostly get what they want over the various networks in the market today. However, as more of that connectivity functionality that resides in the ‘cloud’, increases in device performance are necessary to keep pace with the rich features that reside in the cloud. This puts pressure on SoC design and SoC architectures. Cloud-scale devices are driving SoC complexity due to the following market demands.

John Koeter, the Vice President of Marketing, for the Solutions Group at Synopsys, Inc., gave a presentation at the Semico Impact Conference:  Focus on the IP Ecosystem on May 16, 2012 that looked at the changes occurring in the market today caused by the rise in mobile devices. First, he started of by giving some metrics on the mobile market captured by Cisco.

Semico held its Impact conference: Focus on the IP Ecosystem at the DoubleTree Hotel in San Jose, CA on May 16^th,2012.  The day before, on May 15^th, Cadence Design Systems announced its first IP Subsystem product, a Storage IP Subsystem based around the new NVMe (Non Volatile Memory express) interface standard for flash-based storage applications completed in March, 2011.  This IP Subsystem provides a complete HW / SW verified solution and maximizes the command throughput for interaction between system software and the storage system.

It is intended for those applications that want to replace rotating media in a system with flash-based Solid State Drives (SSD). Vishal Kapoor, Vice President, Marketing Design IP and Services at Cadence gave a presentation at the Semico conference the very next day. He provided some fascinating insights into the forces at work behind the creation of their Storage IP Subsystem.

SoCs are taking to the Cloud.  Just as microprocessors drove the PC boom 20 years ago and the internet drove the communications boom 10 years ago, SoCs are revolutionizing consumer electronics today.  Each boom brought us new applications, rapid decline in product cost, and many more users.  While microprocessors drove computer volumes in the millions, complex SoCs are driving consumer products in the billions.

What is driving SoC complexity?  This was the focus of Grant Pierce’s talk at the Semico Impact Conference on May 16^th, 2012.  As we all know, today’s consumers are looking for higher quality at lower prices.   They want video, voice, data, and audio in everything.  All this convergence pushes the need for multi-GHz performance.

Apps run on everything and apps need “Giga’s” whether it be a 1-3 GHz multi-core CPU, 100+ GFLOPS multi-core GPU, or a 15-50 GB/sec DRAM chip.  And as we use more apps, we will continue to need even more “Giga’s” in the future.  But all these “Giga’s” burn more and more power which is why our devices die so quickly.  These SoCs have gotten so powerful, that today’s batteries can’t afford to power them all at once.

On Wed May 16, 2012 Warren East, CEO of ARM, gave the opening keynote address at Semico Research Impact Conference: The IP Ecosystem.  Mr. East addressed the issue of IP (Intellectual Property) evolution and growth.  Looking forward there has to be more collaboration in the IP world, both horizontally and vertically.

IP companies need to work closely with chip vendors.  Semiconductor companies want to hold onto what they see as differentiators.  But this may work against lowering development costs.

In the last 20 years ARM has enabled wireless mobility.  The fabless model has lowered costs and spawned the growth of many companies.

IP has value.  Mr. East cited Semico’s projection of $3 Billion in projected IP revenue in 2012, a 25% growth rate.  There are over 100 blocks per chip.

There are increasing demands on chip design.  Today the chip is the system.  The board complexity has moved onto the chip.  Designers have to balance power with performance.  There are optimized processing units designed for specific tasks.

Designers at large IDMs first created their own IP Subsystems out of the discrete IP blocks they were already licensing or had developed internally.  This was done to both reduce the level of effort they were expending to create a certain level of functionality and to increase the performance of these discrete blocks in the design. This process has been occurring for the last 4 – 5 years as a captive activity at several of the larger IDMs around the world.

Now, for the first time, major IP Vendors are starting to offer their own IP Subsystems as products available from the 3^rd Party market to designers at companies of all sizes. The IP Subsystem as a product is the embodiment of the desire by silicon designers to move up a layer of abstraction and to design with system-level functions instead of licensing many discrete IP blocks and aggregating them into the system-level functionality they need for their silicon solutions. By approaching the design flow in this manner, a great deal of time and cost can be removed from the design effort. There are several potential benefits to designers in adopting this approach:

Vishal Kapoor was one of the speakers at the Semico IMPACT conference last week.  The key take away is that we need a change in perspective in the IP industry, one that is system driven.  We used to view IP as a “This is what I do” type of product.  But we need that to change to a “What can I do for you?” type of product.

We are in a evolutionary industry.  Around 2000 – 2005, we entered the age of consumerization.  Integration became important because Time to Volume and Time to Market are the two factors that most determine success in the consumer arena.

And what made the age of consumerization possible are standards.  Like the USB standard.  It created a megastore type of environment for IP.  A designer can walk down the aisle and find convenient and dependable parts that can be configured and integrated.

Complexity, though, hit a point when people could no longer do it themselves.  They had to start looking for others who had products that could work with their own.  That point was around 100 blocks.

The evolution of the IP system means that we can no longer just buy single blocks and then wait to see if they all work together.

There are a few questions you should ask to determine quality:

1)      Does it work in the system?

2)      Does it conform to parameters?

3)      Is it configurable?

4)      Does it work in both low and high end products?

Mahesh Tirupattur, Executive Vice President of Analog Bits, led the panel "Who Wins in the IP Ecosystem?" during the Semico Impact 2012.  Other members of the panel were Suk Lee of TSMC, Jean-Marie Brunet of Mentor Graphics, Tony Stelliga of Intersil, and Dr. Naveed Sherwani of Open-Silicon.

Mr. Tirupattur began the panel by describing today's industry as a collaborative model, where customers tell the IP companies what they want, and what they can do better.  What the industry needs is a dynamic ecosystem that improves efficiency, with no waste or redundancy.  It's not just who wins in the ecosystem, but how do we get paid for the value we bring to the table?

If you cover a long enough time period, the small ups and downs of a graphed line can look very smooth.  Semico’s semiconductor wafer demand data goes back to 1991.  When graphed from 1991 to 2016, wafer growth appears to have a very steady upward trend with only a few minor interruptions.  Wafer demand grows at a compound annual growth rate of about 8-9%.  When graphed against semiconductor units it looks very tame.  Looking at Figure 1 you may say to yourself, “The forecasting business can’t be that difficult.”  What’s so difficult about forecasting wafer demand?  Looking at the first graph, one might conclude that capital investment in fab capacity should be very successful if a company stuck with an 8-10% investment rate every year.

Figure 1: Semiconductor Unit Sales and Wafer Demand

Source: Semico Research Corp. & SIA

Reality sinks in when we look at the percent change in wafer demand since 1991.  (See Figure 2)  Wafer demand growth is anything but stable.  There is almost a 50-point spread between the peak growth year to the worst year of decline.  Even if we throw out the worst year (-23% decline in 2001) and the best year (25% growth in 1995) the change in wafer demand growth rates range from -2.7% up to 24.9%.

Figure 2: Percent Change in Wafer Demand

Source: Semico Research Corp. & SIA