Results Analysis

We know that a stock GTX 480 flattens most of our tests, especially at resolutions less than 2560 x 1600, so we expected the Zotac card to do the same.

We weren’t disappointed with the results the card showed in our Dirt 2 benchmarks with the results actually appearing CPU limited at 1,680 x 1,050 and 1,920 x 1,200, even with 4xAA applied. Above these resolutions the Zotac pulled out a small 7-10 per cent lead over a stock clocked card but realistically a stock card is more than adequate for playing Dirt 2 at almost any resolution with lashings of AA thrown in.

Our STALKER: Call of Pripyat tests are more of a challenge, especially for Nvidia cards that seem to struggle with the heavily ATI optimized game. Struggle may be a slight overstatement though as the AMP! card still gave playable results in all our STALKER tests. Again it had a slender lead of 1-3fps over a stock clocked GTX 480 but both Nvidia cards lagged behind the cheaper ATI HD 5870 based cards in this test.

Our Crysis tests are still some of the most demanding tests we run for graphics cards and even the Zotac card had to drop the smarmy swagger it had adopted for our earlier tests in favour of rolling its sleeves up and doing some real work for a change. The AMP! returned good results across the board, performing particularly well when AA was enabled.


We also saw a more consistent difference between the AMP! and the stock GTX 480 during our Crysis testing with an improvement of 8 per cent in both the minimum and average frame rates at 1,680 x 1,050 with or without 4xAA applied. These gains are of particular benefit given that the performance of a stock GTX 480 is just hovering below what we consider playable at 1,680 x 1,050 with 4xAA and the extra few fps from the AMP! really make a difference.

The Fermi architecture again showed its quality when processing anti aliasing during our Bad Company 2 testing as it opened out a good lead over its Cypress based ATI rivals when AA was enabled. Honours were more even when AA was turned off but the AMP! still returned playable results in all resolutions with or without 4xAA enabled, which is an good result. The gap between the AMP! card and the stock GTX 480 was minimal at lower resolutions but stretched to between 8 and 10 per cent at 2,569 x 1,600.

Power results were broadly in line with our expectations, we know Fermi is a thirsty chip and we weren’t surprised to see a slight bump in power consumption over a stock card due to the overclock. Thermal testing on the other hand was more interesting as the Zalman cooler attached to the Zotac card showed its pedigree by knocking a massive 32^oC off the delta T of the stock card which is an excellent result. It’s disappointing that a cooler that has such evidently excellent cooling abilities is marred by a poorly designed and fitted shroud which serves little purpose.

What was also impressive is that the cooler was whisper quiet during all our testing, despite being a great deal more effective than the noisy stock cooler. This is a major positive and makes the Zotac card much easier to live with on a day to day basis as the whooshing sound of a stock cooler can become annoying over time.


Despite the card being pre-overclocked we were eager to see how far we could stretch the Zotac card. We downloaded the latest version of Zotac’s overclocking tool FireStorm (the version on the disk was out dated and didn’t recognize our card) and set about trying to melt our test rig.

Firstly we should say that we were pleasantly surprised by FireStorm - it’s streamlined and elegant, unlike many manufacturer's overclocking software, and recovered from failed overclocks well. It also helped us to push our card way beyond what we thought we were likely to manage and we ended up with a core clock of 840MHz, a shader clock of 1,680MHz and a memory clock of 3.95GHz.

These clocked added an additional 10 per cent to average and crucially minimum frame rates in our [i]Crysis/[i] tests which is impressive given the card already out performed a stock card by a similar margin. This should also be encouraging for card manufactures looking to produce ‘super overclocked’ GTX 480s and almost makes the overclock that the Zotac came shipped with seem conservative.

Conclusion

The Zotac GTX 480 AMP! is the first pre-overclocked GTX 480 we’ve seen. We were sceptical about how impressive such cards were likely to be, given how hot and loud the stock card runs.

Happily we can say that the Zotac has proved us wrong to a degree; with the right cooler a GTX 480 can be pleasantly quiet. The faulty shroud (which may not be a problem with all the VF3000 coolers but certainly was with the two we have in the labs) meant we didn't get off to a great start with the card, and you really shouldn't have to remove parts of a £510 graphics card to make it work correctly. That said, a few weeks or months into owning this card the shroud would long be forgotten and you'd still be enjoying a card that's dramatically quieter and cooler than the stock model.

The AMP! also has its factory overclock and extra overclocking ability in its favour. Perhaps other GTX 480s could be pushed to such a level with a good third party cooler fitted, but if you buy the Zotac you don’t have to fit the cooler yourself and you get a warranty on the overclock. You pay for this though - you'd have £50 left over if you bought a stock GTX and a VF3000 separately.

Don’t get us wrong though, the AMP! is a great card and it does what it can to address the failings of the original GTX 480 by being cool-running, quiet and quick. It can't improve on Fermi's outrageous power usage and of course, it can't help the price, either. At £500, it's a luxury product with a not entirely sensible price.

• Performance
• x
• x
• x
• x
• x
• x
• x
• x
• x
• -
• 9/10

• Features
• x
• x
• x
• x
• x
• x
• x
• x
• -
• -
• 8/10

• Value
• x
• x
• x
• x
• x
• -
• -
• -
• -
• -
• 5/10

• Overall
• x
• x
• x
• x
• x
• x
• x
• -
• -
• -
• 7/10

Score Guide