Tony Massimini's blog

June 21, 2010

There was a lot of excitement last week at the E3 Show in Los Angeles.  There were numerous announcements, demos and news on new game titles.  On the hardware side what grabbed a good deal of attention were Sony and Microsoft products for motion control to compete with Nintendo Wii.

The console makers were also touting sleek new console designs.  These are slimmer and operate with less noise and lower power consumption, thus generating less heat.  Prices for consoles have come down over the years.  Ironically, the snazzy new peripherals will help make up the difference in price.

However, all of these appear to be about further enhancements and peripheral developments that improve the gaming experience of the current platforms.  3D games will be achieved through SW.  The console makers have taken advantage of improved semiconductor manufacturing technology to reduce the size and power consumption of the CPUs and GPUs.  This has enabled designs that are less expensive, lower power consumption and sleeker designs.

However, the basic design of the CPUs, GPUs and support chips have not changed since their introduction in late 2005 and 2006.  There is no discussion about next generation platform designs, similar to those that preceded the introduction of the current generation.

May 25, 2010

The world of TV is changing rapidly. The recent Semico Spin article by my colleague, Michell Prunty, presents the scenario that is playing out. The entire business model for TV is changing. Over a year ago cable TV companies were reporting that their cable subscriptions were dropping, but their internet business was growing. Clearly, people like Michell were contributing to this trend.

Over the course of the last decade there has been strong growth for TV set-top boxes as well as variations, such as IP-TV and satellite TV. Concurrent with this has been the transition to digital TV and HDTV. Consumer electronics companies are now working on 3DTV. But the new business model presented by Google TV has technical implications as well.

Google’s Android is open source. In terms of a processor it is considered architecture-neutral. It should be noted that the architecture with the dominant market share in DTV, TV STB, etc. is MIPS. The company has worked closely with Google to develop and support a software ecosystem for Android on MIPS-based solutions.

The digital home market is an important market for MIPS. It will work hard to maintain its dominant position. Since Android is processor agnostic there will be increased competition from ARM, Power PC and Intel’s Atom. The new platforms will have to deliver the performance and independence that consumers are looking for.

The long awaited Moorestown platform has been released by Intel. This was presented at the Intel Developer’s Conference in September 2009. The Moorestown platform consists of the new Z6xx Atom CPU (formerly known as Lincroft), a controller hub (formerly known as Langwell), a mixed signal IC and wireless options.

The mixed signal IC (MSIC) integrates several analog and power functions. The mixed signal ICs and the wireless options are available from several other IC vendors, not Intel.

What is significant about this announcement is that Intel has delivered on its promise to deliver a highly integrated version of Atom with very low power consumption and high performance. Intel was able to integrate many of the functions that were previously in the chipset into the CPU. These functions are a scalable bus interface and coherency engine, 3D graphics, display controller, video decode/encode and the memory controller. The previous generation Z5xx series (Silverthorne) CPU and this latest version are both manufactured at the 45nm node. The Z6xx has an Ultra-Low Power (ULP) core.

Intel is touting its new OS Power Management (OSPM). This is a power management program aimed at the entire platform not just the CPU. The OSPM provides fine grain power management to the CPU, controller hub and the MSIC.

The key components of the Moorestown platform are available today. Pricing was not provided.

Oh it has been a busy week of rumors and speculation in our industry.

First there was the report that Apple is looking at designing in AMD MPUs in future MacBook designs. This should come as no surprise since AMD MPUs are 80x86 and Apple’s PC product line is based on Intel MPUs. Apple is the only major PC vendor that does not offer products with AMD MPUs. I am surprised that it has taken this long for Apple to consider AMD as a second source.

However, this set off speculation that Apple would buy AMD? That makes no sense. Apple would not cut itself off from a major MPU vendor like Intel. Also, there would be legal issues about an OEM buying a key chip vendor that supplies its competitors. How this idea ever made it to print or on-line media is beyond me. I guess it is one way to get attention.

We then had the rumor that came out of England that Apple was considering buying ARM. It was reasoned that Apple was flush with cash, around $40 billion and that ARM looks so attractive. Acquiring ARM could cost about $8 billion. This one makes no sense for so many reasons.

Apple is not the kind of company to make big acquisitions. It goes out and finds small strategic companies and integrates the talent and technologies in the process of building up its own resources. The acquisition of PA Semi for the design team that came up with the iPad A4 processor is a recent example.

The Apple iPad has hit the market, at least in the US. It has been shipping well, though not as high as some had predicted. Incidentally, Semico has a conservative and slow roll out scenario for the iPad with sales picking up at the end of 2010 and early 2011.

Apple delivered the first iPads starting Saturday April 3, 2010. At the end of the first weekend about 300,000 units had shipped. Sounds impressive, but around 90% of these were pre-ordered in the weeks leading up to this launch. The huge lines at Apple stores and Best Buy did not materialize. Some estimates had foreseen over 700,000. By the end of the week around 450,000 had shipped. Apparently the early adopters came out in force for Apple. Will they keep the ball rolling? This is not to denigrate the iPad’s early shipments. Rather it shows how hype can drive some rather unrealistic expectations.

It has been observed that iPad’s early success matches iPhone’s initial sales. Some history; the iPhone was introduced in 2007 and sold 1.4 million in the first 6 months. It reached 6.1 million by the end of its first year. The iPod was introduced in 2001 and sold 372,000 in the first year. After opening the iTunes store, the iPod shipped over 1 million in its second year. Both products were introduced into existing markets. Apple was offering innovation to product categories that consumers were already familiar with and was able to take market share.

The Apple iPad will be shipping April 3, 2010. If you have already pre-ordered, you will get it at that time. If you still need to order the iPad, you will be waiting until after April 12. According to recent news items, the pre-order levels may have exceeded Apple’s initial expectations.

Apple experienced a huge surge in orders the first day pre-orders were available. Pre-orders have fallen to a slow but steady pace ever since. After two weeks of orders it appears that approximately 240K pre-orders are on the books. Now the question is not so much whether Apple can deliver the product on time, but can it deliver on the hype?

There are high expectations for the iPad, but is the hype just Apple’s loyal following? What will Apple offer to make the iPad attractive to the mass market? How will the iPad compete with e-readers and netbooks? What is the market potential for tablet PCs? What are some of the issues for the semiconductor industry and what are the business opportunities for chip vendors?

Semico Research has prepared a market brief, “Apple iPad: A New Computing Paradigm or the iPhone on Steroids?” Semico examines various questions and issues around the iPad and tablet PCs in general. The market segment of tablet PCs has yet to truly emerge. This market brief will address some of the competitive factors that are likely to influence its development.

About this time last year not many people wanted to believe me when I said that the Micro Logic market (microprocessors, microcontrollers, and DSP) would bounce back in 2H 2009 and with a strong upward trend.  Fear, uncertainty and dread (FUD) due to the worldwide economic crisis was driving companies to make severe cutbacks on their manufacturing plans and consequently reduce inventory levels.  Some plans, like Dell and HP lowering production to unusually low levels for 1Q 2009 were self fulfilling prophecies.  It just appeared that the steps taken by most OEMs were extremely drastic.  Semiconductor inventory levels had not been that low in anyone’s recent memory. 

The automotive market had lots of problems of overcapacity and high inventory levels of vehicles.  This is the largest market for MCU sales.  Governments stepped in with programs to stimulate consumption.  The US passed the “Cash for Clunkers” program and Germany had a similar one.  As we approach the end of 2009, vehicle sales are improving in all global markets.  Semico’s discussions with key suppliers to the automotive market indicate a more positive outlook at this time.

Connectivity has been a driving force in electronics.  It is not just important for computing and consumer applications.  The industrial and medical markets show increasing need for connectivity.

Microchip Technology recently announced new members of its PIC32 32-bit MCU family to address this growing need.  Increasing connectivity for embedded control applications puts an increasing demand on memory requirements.  These communication protocols need more software stacks that must run simultaneously.  The new PIC32MX5/6/7 families are designed specifically for these data-intensive applications.

Microchip’s PIC32, a MIPS based product line, now offers memory sizes ranging starting at 256KB for Flash and 64KB for RAM, going up to a combination of 512KB and 128KB for Flash and RAM, respectively.  The company is integrating a wide variety of communication peripherals in different combinations: Ethernet, CAN and USB modules, along with UARTs, SPI, I2C and DMA channels.  The MCU is based on a high performance MIPS core at 80MHz.

Some of Microchip’s target markets are:

Last week (August 28, 2009) Intel raised its revenue forecast for 3Q 2009.  The company stated that due to “stronger than expected demand” for microprocessors and chipsets it has raised is revenue forecast to $9.0 B, plus or minus $200 million.  The previous outlook given by Intel in July was between $8.1 and $8.9 billion.

Intel is one of several other high profile companies in high tech who have recently indicated seeing signs of a recovery.  Dell had lower profits than a year ago, but not as bad as had been expected.  Dell is anticipating improving sales in 2H 2009.  Hewlett-Packard has seen its profits slide also, but the company has said that business is stabilizing.

Semico Research has forecasted since early 2009 that the semiconductor industry would see a strong recovery starting in Q2 2009.  This has been reported in several of our publications, including the IPI Report.  Semico’s PC forecast and related MPU forecast has changed very little since early 2009.  The recent news, including Intel’s revised 3Q 2009 outlook, supports Semico’s views.

Intel’s MPU business usually represents about three-quarters of its total revenue stream.  Historically, Intel holds just over 80% of the MPU market, per the SIA/WSTS Blue Book.  Therefore, Intel’s revenue prediction is in line with Semico’s MPU forecast for 3Q 2009 of $8.2 billion.