Nvidia's original GeForce GTX 580 is a lightening fast card, and is the fastest thing on single GPU legs right now. So an overclocked version, like this GTX 580 AMP! from Zotac is surely going to be a beast.

Buying a pre-overclocked graphics card though is like buying a car with all the options.

You might not see a major performance gain, but you're safe in the knowledge that, in its class, your card is probably the best.

The flipside, of course, is that pre-overclocked cards play on this perceived value, and the performance gain is rarely in line with resultant price-hike percentage over stock.

These are margin-making products for board partners, and offer flagship performance in-class that influences the way we feel about the rest of the company's line-up.

This is no great conspiracy; it's intelligent brand-management and it's been going on for years.

Besides, building a custom board-and-BIOS platform in which to plop a GPU takes time, effort, and ultimately money. When you're dealing with a GPU as complex and intricate as the GF110, you have to take time to balance tolerances and get this stuff right.

The already blistering GF110 isn't noted for its overclocking headroom, and Zotac's GeForce GTX 580 AMP! offers noticeably – though not vastly – higher component-speeds than the stock GF110-based GTX 580.

Up from 772MHz to 815MHz (a 5.57% gain) on the core and from 4008MHz to 4100MHz (a 2.3% gain) on the memory, to be precise.

The cooler is the standard GTX 580 reference unit – which in itself is anything but standard, with its innovative vapour-chamber design.

Price-wise, you're looking at 16% premium over the cheapest stock GTX 580 we can find, against considerably lower performance gains.

So is it really worth the extra outlay over a stock card?