Kicking off the Intel Developer Forum 2005, Craig Barrett outlines where Intel is and where it's heading.
Day one of IDF 2005 kicked off with a keynote speech from Craig Barrett – Chief Executive Officer at Intel. Craig was keen to tell us that digital electronics has dominated the past 50 years, and that the transistor has been responsible for nearly every technological innovation during that time. Now, it comes as no surprise that a chip manufacturer will credit the transistor with such importance, but to be honest, it’s probably a pretty fair statement.
Citing Moore’s law, Craig said that Intel is continuing to double the amount of transistors in a chip every 18 – 24 months, but it has reached a point where the company has had to start thinking differently when it comes to increasing processing power. As such, Intel is trying to move away from the constantly increasing clock frequencies as a method of improving performance, and instead is pushing the concept of dual and multi-core processors.
Craig didn’t delve too deeply into the multi-core strategy, since that was going to be covered elsewhere. However he did touch on the roadmap for the CPU manufacturing process. This year (2005) we’ll see the introduction of a 65nm process, while in 2007 this will drop to 45nm. In 2009 we should see 32nm chips, while in 2011 Intel plans to have reached a 22nm process.
Craig claimed that Intel should be able to go all the way down to around 5nm, before it hits real manufacturing problems. In fact, in a previous briefing Intel stated that the ultimate limit for silicon is actually 1.6nm, but once you get down to 5nm leakage becomes such an issue that it’s not commercially viable to go any further. The main point being made though, is that there’s still a lot of life left in silicon.
But Barrett also wanted to touch on how important mobile technology is. There’s no denying that Intel has given a huge boost to the wireless networking initiative with its Centrino platform, but as Barrett pointed out, WiFi is too limiting a technology. Intel feels that the answer to this problem is WiMAX and it is fully behind the implementation of WiMAX infrastructures around the world.
Barrett clearly wants to see widespread, high-speed wireless Internet access, with greatly increased bandwidth, and I don’t think there’s a technology enthusiast out there who doesn’t want the same thing. With more and more data distribution taking place over the Internet, bandwidth needs to increase massively – to quote Barrett “2Mbit isn’t broadband, 50Mbit is broadband” – Amen to that!