It seems that the two companies, who have been fighting for GPU market supremacy over the last years, have picked diametrically opposed strategies for reaching their goals.

Both companies have recently made public their latest graphics chips, but the two models are very different, as well as the architecture that they will fit in.

If Nvidia goes for the bigger, better, all on one chip paradigm, AMD hopes that a simple, power efficient and scalable design will bring the company again on the first place on the market.

Nvidia’s GTX 200 series chips manage to cram more than 1.4 billion transistors on the same silicon strip, and features 240 processing units. But everything has a price: it has a much larger size than its predecessors and needs about 182 watts for the GTX 260 and 233 watts for the GTX 280 to operate.

Performance-wise, Nvidia says that the GTX series chips are able to make an enormous amount of calculations, which makes them able to render the simultaneous movement of more objects during game play. What is more, the company says that converting a high definition movie into the format that is playable on an iPod takes about 35 minutes, compared with the 5 hours that the same process takes on a regular system.

However, their bulky form factors make these Nvidia products unsuitable for notebooks.

On the other hand, AMD’s new chip, which will be embedded in the HD 4850 and HD 4870 graphical processing units, crams less than 1 billion transistors, but has the capability to be used in a multicore architecture. What this means is that AMD will be able to address the midrange segment of the market with a graphical processing unit that will embed one chip, while using the same type of chip, but in a multicore architecture to address high-end demands.