Intel and Larrabee - Is it Time to Start Talks with nVidia?

Larrabee, as most power users know, is the ‘answer’ to the question “Why does Intel always make graphics that suck?” It is to be the Core to AMD’s Barcelona, in the war of the big guns in computing.

Anyone who has used onboard Intel graphics, or was unlucky enough to purchase one of the older i740 cards, knows that Intel graphics are more than an order of magnitude behind midrange designs by discrete graphics chips manufacturers.

Now an in-depth article on Bright Side of News gives the long, winding, and fraught with inept changes, description of the evolution of the Larrabee chip and the problems that Intel has had bringing it to market.

It would be a complete disservice to the writer of the article, to bring bits of it here, as it needs to be ingested whole, to get the full effect.

But, if I can try to sum it up in a few sentences.

Intel has known it needs to have an answer to the market need for discrete graphics cards, because, with the advent of operating systems that are moving the bar higher, nearly everyone will need more power than current Intel onboard graphics can provide.

More than that, it needs to have a way to stave off the number of manufacturers that will be moving to the nVidia Ion platform, because that gives Intel only money for the CPU, and not the entire board (graphics and glue chips). Intel also needs to bring a powerful graphics core to market, that can be used in its CPU+GPU offering, to compete with the AMD Fusion project.

Also, Intel needs to have a graphics solution for that part of the market that will always want a discrete graphics card, because of gaming, and other high powered rendering. Here it is not entirely about money; it is also about pride, because the high end not only points the direction of the entire market, it tends to make the statement that if one company has the best at the top, it makes sense that at each lower step, that same company would excel. (though most know this is not necessarily true, it holds for the unwashed public)

The perception is reality situation comes into play here – in the end, if Intel, that was expected to come out a couple of years ago, with a card that would trounce ATi and nVidia, in the manner of the smackdown that was the first Core processors, can’t deliver, then what has happened? Consumer confidence is a very big thing. It makes Intel the corporate choice, simply based on that confidence. If things were done rationally, based on cost analysis, AMD would have a much larger corporate share. But that is not what happens.

So much is on the line for Intel. It must make the decision to get on with it, or give up and buy technology, in much the same way Microsoft buys what it cannot seem to do itself. It has worked for Microsoft, so perhaps it could for Intel.

The BSN article states that Intel must invent, rather than buy, for that is the expectation. I think that a purchase of nVidia would be a sure win, and in the crap shoot of reality, is a much safer bet.

§

⌘

⌘

•