Larrabee might not be that great...

INVESTMENT BANKING BOFFIN, Ashok Kumar, an analyst at Collins Stewart, reckons Intel’s upcoming Larrabee graphics chippery has an uphill battle ahead if it wants to play with the big graphics boys.

Kumar reckons Intel is being either supremely daring or supremely hubristic in proclaiming its Larrabee will be able to outperform Nvidia’s best GPUs, noting, "Nvidia has had over 10 years to optimize the 3D graphics pipeline, the necessary drivers, the platform connections needed to supply the memory bandwidth required, and to work with the software and apps developers".

But Kumar may simply be playing it cautious. After all, Intel has a knack of dominating every other chip market it tries its hand at, so why not give standalone "discrete" graphics processors a go?

The chip giant has already carved its name deep into the integrated graphics silicon market, but admittedly this doesn’t mean much when it comes to discrete graphics, which are highly complex chips, chock full of over a billion transistors and hundreds of stream processors.

Also, with Larrabee’s expected release in 2009-10, Intel will be muscling its way onto a playing field which, up until now, was territory reserved only for Nvidia and AMD's ATI.

Intel, however, don’t seem to be paying the competition too much mind, with executive VP and chief sales and marketing officer, Sean Maloney, saying in an interview last week "I certainly expect Nvidia and ATI to carry on being successful," and adding “We're trying a different approach”.

It is precisely this “different approach” Kumar seems to be taking to task in his report, with Intel using software which will have to run properly on multiple processing cores and based on what Kumar calls an "antiquated" Pentium design.

Larry Seiler, the senior principal engineer at Intel's Visual Computing Group, seems sure of himself though. According to Seiler, Larrabee's Pentium-derived design "has five times as many cores, each core has a vector (processor) unit that is four times as wide. So for throughput computing, potentially it can run twenty times faster".

Seiler isn’t the only Intel employee who seems to be barely restraining his enthusiasm. Pat Gelsinger also appears psyched up about Larrabee’s prospects, especially as it will plug into future Intel platforms like Nehalem, Westmere and Sandy Bridge.

Kumar, however, remains skeptical, noting "If Larrabee ends up knocking out Nvidia, it will be a shocking upset", especially, he adds due to the "high efficiency of Nvidia's existing designs".

Still, even if Larrabee only manages to be just a little bit successful, this can only be bad news for both AT and Nvidia, because any kind of Intel gain in the graphics market will surely be their loss.