Larrabee derives most of its number-crunching power from these vector units.

They employ dedicated processors, called vector units, that can rapidly perform long strings of calculations.

This is accomplished by having two 256 bit vector processing units in each core.

The vector unit performs add/shift, multiply, divide and logical operations.

For additional performance the vector units supported a special multiply-and-add instruction that could retire two results per clock cycle.

Another difference is that the main scalar units of the processor ran at half the speed of the vector unit.

This forced most programs to hit the high setup cost of the vector units, and generally the ones that did "work" were extreme examples.

In the G4, up to eight simultaneous operations can execute in a single clock cycle in the vector unit.

Then there's this otherwise very confused PC World article, where the writer obviously heard something about a 512-bit vector unit.

In fact, from what I've heard, very few developers are using both vector units.