By its own lofty standards it’s been a tricky couple of years for chip giant Intel. Bitter rival AMD stole the CPU performance crown with products like the Athlon 64, major customers like Dell started producing AMD PCs and even the once ubiquitous ‘ding, ding, ding, DING!’ signature tune of ‘Intel inside’ appeared to be being phased out.
While the demise of Intel’s theme tune was not universally mourned, it was, however, greatly exaggerated. In the past year Intel is generally accepted to have retaken the performance crown from AMD with its Core 2 brand of CPUs. It’s possible that the process of absorbing ATI will have caused AMD to take its eye off the ball, but Intel still had to make a move.
To reflect on this pivotal time and find out how Intel intends to cement its renaissance I meet Intel’s digital home business director Steve Shakespeare and start by asking him about the last 12 months. “We were facing some challenges a year ago,” he says. “We had a great product in the form of Centrino, we’d just launched Viiv as our digital home technology, but frankly we had some challenges on the desktop side, especially concerning power and performance at the top end.
“In the second half of last year we launched Core 2 Duo and have been receiving a great deal of positive feedback on that. 12 months ago we were focusing on the platforms – Centrino and Viiv. What’s evolved is that we’ve effectively released a new engine which we want to ensure is perceived as the common thread through the platform brands and marks a retaking of the technology crown.”
With some trepidation that I ask how Intel retook the technology crown. “It’s been ten years since we had a substantial refresh of our CPU architecture – which is what Core 2 is,” says Shakespeare, “and we’ve realised that we need to accelerate the pace at which we introduce new technology. Consequently, soon after we released dual core CPUs we announced quad core and we’re very much on a multi-core strategy for the business.”
We all remember the gigahertz wars of five to ten years ago. I remember being very proud to be the owner of a 1 gigahertz (AMD) processor back in 2000, but when clock speeds got to around three gigahertz the unthinkable happened: clock speed stopped being the primary measurement of the desirability of a processor. Suddenly consumers couldn’t just think ‘the higher the number the better the processor’. Mercifully we are now entering the era of multiple cores and I ask Shakespeare if cores will be the new gigahertz.
“The direct correlation between clock speed and performance went away a few years ago as microarchitecture changed and suddenly Gigahertz were not a meaningful way of differentiating one chip from another,” he says. “We then saw the start of what ultimately became the multicore strategy. You saw us use hyperthreading, which was like one and a half cores – using a few spare cycles. We also redesigned the innards of the PC, thus changing the whole profile of Gigahertz needed to run.
“It’s apparent that the number of cores could be the new vector for performance in the marketplace. To take advantage of many cores applications will increasingly be written to take advantage of multitasking. So clearly the number of cores will become an important factor in decision making about performance products.”
There is another measurement associated with processors and that is nanometres. Stretching the boundaries of my technical awareness I discover that this refers to the width of the gate within a transistor and that the smaller this number is, the more transistors you can cram on to a piece of silicon and the greater the functionality you can extract from it. Making smaller and smaller transistors appears to be a central part of Intel’s strategy to stay on top.
“We now have a six year roadmap which involves regularly refreshing our microarchitecture and bringing new products to the market,” says Shakespeare. “For example, we’re currently at 65 nanometre technology and we’re going to do a shrink to 45 nanometre technology in the second half of this year. This will enable us to increase the performance and decrease the power consumption of the products.
“If you can shrink the product you can add multiple cores to it; we already have quad core and we will go beyond that. The more transistors we can get onto a die, the more functionality you can put into the product.”
So clearly it’s all about cores now, then. Sadly not; as with so much in life it’s a bit more complicated, as Shakespeare explains. “The vector of the number of cores will become an important aspect of CPU evolution, but it won’t be cores alone. So, for example, this May we launch the next generation of Centrino platform – codenamed Santa Rosa. There will be two brands: Centrino Duo and Centrino Pro – the business platform – which is basically VPro for Centrino.
“As well as some performance improvements we’re also bringing in some things that are really relevant to the consumer, like the 802.11n wireless standard, which is expected to be ratified in the first half of 2008. Someone using Centrino with an ‘n’ access point will have up to five times faster wireless performance, better range and will work better through walls.
“There’s a piece of technology that’s coming with it – codenamed Robson – that will be called Intel Turbo Memory. This is on-board memory that will be used to get the machine out of hibernation very quickly and will also accelerate a number of applications. This should speed up the overall user experience and will be an optional installation with Santa Rosa.”
Before we forget, Intel does have another platform out there – Viiv. My impression is that it hasn’t had the impact Centrino did, but reason is mainly down to the more complex message. Intel succeeded in making people think ‘Centrino’ when they were buying a laptop, but what is the product they intimately associate Viiv with? A media centre PC? If so, the market for those is far smaller than the laptop market, which would explain the relative quietness of the brand.
Shakespeare reckons there’s a lot more to the Viiv market than just media centre PCs, however. “Something in the order of 15-20 per cent of the desktop market is now Viiv PCs, that’s a faster growth rate from launch than Centrino experienced,” he says. “The consumer electronics form factor (i.e. set top box) is a relatively small fraction of that; roughly five per cent. What we’re trying to say is ‘the PC doesn’t have to be in your living room’. Through DMAs and LANs people can access their PCs and hence their content, without the need for it so be in the same room as them.”
Intel’s platforms are made possible by it providing more than just the CPU. An Intel platform in its most basic form consists of a processor, a chipset and software. One problem Intel chipsets face, however, especially when it comes to graphics, is that they’re perceived as very much an entry level graphics option.
This, combined with AMD’s take-over of ATI and the launch of graphics-heavy Vista, has led to speculation that Intel may be considering dabbling in the discrete graphics market. I ask Shakespeare to clarify the situation. “Intel is actually the largest embedded graphics manufacturer in the world,” he says. “We believe there is demand for our platforms and will continue to invest heavily in them.
“Regarding graphics, we already have chipsets (945 and above) that run Vista very well indeed. It’s not the case that you need a discrete graphics card to run Vista well; you can run Vista in Aero mode on any 945 chipset. There is a lot of confusion in the marketplace about what you need to run Vista. Of course we will continue to evolve our embedded graphics solution. The discrete graphics manufacturers have clearly demonstrated that they can add value, particularly when it comes to things like high-end gaming. There will always be a marketplace for specialist products to supplement the core CPU functionality.”
Shakespeare seems to be saying Intel is happy to leave the hardcore gamer niche to the specialists, while it strengthens its position with the mainstream, but I might be wrong. Only time will tell. To conclude I ask Shakespeare if he has any message for independent retailers and if he can give PC Retail readers an idea of what we can expect from Intel in the future.
“As I see it there are something in the order of 600 independents in the UK that are successful and I strongly believe they can continue to be so,” he says. “Intel’s Viiv and Centrino platforms already offer help to retailers in selling solutions and there’s plenty more to come. Intel’s got a separate division focused on digital health. It’s very early days but we think there is an opportunity over the long- term to offer personal healthcare solutions, to over guidance to users on measuring their health and fitness using computer technology.
“Another area we’re looking in is what we now call Ultra Mobile Devices (UMD). One use for these is location based services. For example if you’re stood in the street and you want to find out where the nearest café is, the UMD tells you where it is. The UMD would have a constant broadband connection and be delivering content all the time. To our mind a UMD is a GPS system on steroids that is always broadband connected. These would be much smaller than the early UMPCS, like the Samsung Q1, and much more powerful. WiMax is a big thing in this area.”
If I had to take just one thing from our conversation it is that Intel is heavily involved in a large proportion of the technology market and is working harder than ever to ensure that remains the case. A key to achieving this is to take very technical products and turn them into concepts your average consumer can understand and desire. Intel’s return to form in this respect promotes technology as a whole and thus benefits everyone in the industry. AMD might have something to say about that, however.