«»
And what’s behind our ratings
As always, we use real-world games to test graphics cards rather than synthetic benchmarks, as real games provide more meaningful results that are simple to interpret. We’ve chosen two titles that accurately tell us how a card will handle both DirectX and OpenGL games, and we run a range of timedemo tests for each.
Despite being reasonably old, Far Cry is still a tough test today thanks to the latest patch, which adds demanding HDR (high dynamic range) lighting effects. These provide far greater contrast ratios than before, allowing for extremes of light and dark along with superior vibrancy and detail. The impact is nothing short of spectacular and, once you’ve experienced HDR in action, you won’t want to switch it off.
We’ve chosen Call of Duty 2 because it really pushes cards to their limits.
As well as being a good test of OpenGL optimisation, it’s extremely texture intensive but also very scalable in its video options.
Every card on test is capable of running recent games if you’re willing to drop to 800 x 600 and lower the detail settings. But while some gamers will be happy with this level of detail, others will want to experience the latest games in their full high-resolution glory. This is why our toughest tests have all the detail options set to Very High with HDR enabled in Far Cry, and the maximum possible quality in Call of Duty 2.
The entry-level cards however, are tested at 1024 x 768 with AA (anti-aliasing) and AF (anisotropic filtering) disabled, and with Textures set to Normal in Call of Duty 2. For the mid-range cards ($200 to $410), we run these 1024 x 768 tests in addition to a second set at 1280 x 1024. In Far Cry, we still leave AA and AF switched off, but Call of Duty 2 is run with 4x AA and 8x AF as well as High Textures.
Finally, the cards costing more than $410 are put through their paces at 1600 x 1200 in addition to the 1280 x 1024 tests. In Call of Duty 2, we select Extra Textures, 4x AA and 8x AF, and in Far Cry we opt for 8x AF. Currently, NVIDIA cards can’t simultaneously perform HDR rendering and AA, so to retain a level playing field between ATI and NVIDIA cards we leave AA turned off. Only the most powerful cards will emerge from this assault course with playable frame rates.
In order that we find each card’s maximum performance, we build a test rig that eliminates all other bottlenecks from the system. To this end, we use an Asus A8N32-SLI Deluxe motherboard, which uses the Nforce4 dual-X16 SLI chipset. We fit a 2.6GHz AMD Athlon FX-60 processor and 1GB of Crucial PC3200 RAM, split over two memory sticks. To ensure the hard disk doesn’t slow things down, we use a 36GB, 10,000rpm Western Digital Raptor.
How we work out the ratings
At the bottom of each review, you’ll see four different star ratings: one for Performance, one for Features, another for Value for Money and a final Overall score. The Performance score is calculated by summing the average frame rates achieved by each card in both games. The result is then converted into a score out of six, which is comparable across the whole group.
For the Features score, we carefully allocate points to each package. We reward the array of outputs and inputs, as well as any supplied cables and applications. Passive heatsinks or quiet fans are also noted, as are games bundles.
The Value for Money rating is based on a weighted average of the Performance and Features & Design scores. The weightings depend on the category: features are weighted higher for entry-level cards, and performance for high-end models. We then factor in exactly how much each card costs (including delivery) to give a bang-per-buck result.
The Overall score is simply a straight average of the other three scores. Rather than giving a Labs Winner award this month, we have given a Recommended award to the best cards in each price category.
This article appeared in the August 2006 issue of PC Authority.
«»
Comments
Be the first to comment on this article.
Thoughts on this article? Add a comment below.
