Sunday, September 23, 2007

A look inside the x86

As usual, last week's Intel Developer Forum was a fire hose of information with several useful droplets here and there buried in a huge flow of generally useless hype and executives who were both amazingly available to the press and well trained about not saying anything substantive to them. The big interconnect bits included the news about USB 3.0, wireless USB 1.1 and PCI Express 3.0 I have blogged about below.

There were two other interconnect tidbits somewhat buried in my EE Times wrap up in print this week. Intel gave a marketing name to what it had been calling CSI. However, the company gave almost no detail about the QuickPath coherent interconnect to be used starting with its Nehalem processors late next year, except that it will gluelessly support two and four-way servers. Analysts think it will be more than just Intel does HyperTransport, four years late. But other than the unauthorized but well detailed white paper from fellow journalist David Kanter, little is publicly known.

Separately, it was interesting to hear from Ajay Bhatt that Intel uses PCI software semantics in its proprietary on-chip interconnects. This is probably one of several Achilles Heels that will make it hard for the world's largest software company to step graceful and late into the world of system-on-chip design. This ain't the stepwise world of the 986 running at 6 GHz anymore. Good luck!

1 comment:

Anonymous said...

The Achilles heel of the entire computer industry is software and the passion against change. The complexity and near inability of the software industry to progress beyond the legacy environments - applies equally to Windows, Linux, Unix, OS X - you name it, they all suffer from it - is what keeps real customer software costs as nearly high as their emotional dissatisfaction. The recent woes plaguing Vista illustrate just how much the legacy mind set dominates technology innovation. People don't believe or perhaps don't care that Microsoft has invested billions to create and deliver a higher quality at the initial launch of a major OS than prior versions such as XP. They quickly forget just how poor the prior version was when first introduced and so they rant all the more about every flaw. All OS suffer in this regard; it is just Windows dominates 90% of the market so their woes garner the most attention.

You observe that it was a surprise that Intel uses PCI as a software model for integrated devices. This should not be a surprise as it has been done for many years well before PCI Express was a glimmer in any one's eye in such common components as south bridge. PCI software model is comprehended by all firmware and OS software so wherever possible the industry looks to use it rather than invent something new. This works well and benefits all when there isn't any new hardware functionality to be exposed through PCI. However it is disingenuous when people assert that there is no new software required for a particular PCI Express feature that requires explicit control to enable or configure. Every bit of such extended functionality equates to a new piece of software.

What should be a surprise is just how often people believe this all comes for free or is completely transparent - people do a great disservice to the industry at large when they don't reveal or gloss over all of the real costs to execute a technology feature or blow it out of proportion with its benefit.

As you noted in one of your other postings, IDF is unfortunately becoming a fire hose of non-substantive information with a lot of people saying a lot without saying anything. I find it difficult to listen to the keynotes in particular because of the scripts - they all seem like a bunch of politicians trying to duck every question of substance. I find it increasingly difficult to sit through many technical presentations because they are being permeated by marketing and becoming largely devoid of substantive technical education. I am also disappointed in how the chalk talks which were seen as a breath of fresh air with real interactive discussions being run by marketing people all too often. If Intel wants IDF to attract developers, they need to move the marketing people out from many of these presentations and turn it over to the technical leads with a focus the technical details instead of visions and strategies. Education should be made the goal of technical tracks while marketing remains confined to the strategy sessions. Keep it honest and real and you'll attract more people.

 
interconnects