Performance Conclusion
Through the previous page of benchmarks and performance discussion, I kept mentioning real world performance rather than what board won each benchmark test. With a performance delta (difference) across every single test being no more than around 3% in either direction, the board that won the benchmark becomes irrelevant. Real world performance is king and while performance analysis is necessary to draw a real world conclusion, too much emphasis on the finer details can lead to uninformed choices about what to buy.In the end, both boards perform identically. We used the sealed box example to illustrate that the 2 boards being compared here, the reviewed AT7 and the previously tested 8K3A, are identical in real world performance and the example would hold up in real life testing.
So we can't fault the AT7's performance. It's a solid KT333 solution with excellent tweaking features for those that want to extract the maximum and a good set of defaults (that were tested here) for out of the box performance on a par with the 8K3A.
The added USB controllers in particular did pose a small cause for concern before starting to benchmark the board since I've found that even when idle, they can suck away performance but that didn't show up here. A sensible implementation of the southbridge ports and the added 2.0 ports from the VT6202.
With increased bridge to bridge bandwidth in forthcoming chipsets, any performance decrease that might have arisen as a result of the extra controllers will disappear as it's swallowed up by the new bridge links (over 1GB/sec in chipsets being released this year!).
Overall it's fast, tweakable and on a par with 8K3A. This makes it the joint quickest AMD DDR solution I've had the pleasure to test. Both boards would also be on a par (within our observed 3% delta) with the MSI KT3 Ultra that Tarinder looked at making all 3 excellent choices.
A quick note about usage issues and then the conclusion!
