Performance in Nominal Mode and During Overclocking

We used Asus P6T mainboard as a reference for performance comparison in all our previous reviews of Intel X58 Express based solutions. Although we didn’t achieve maximum CPU overclocking on this mainboard, it performs pretty fast during overclocking as well as in nominal modes. Take, for instance, the results of our comparisons vs. DFI LanParty JR X58-T3H6 or EVGA X58 SLI Classified, where Asus P6T is convincingly ahead of the competitors. This time we’ll do exactly the same. At first we will compare the performance in nominal mode, when the boards set all parameters themselves.

Overall, these mainboards perform very close, except Everest benchmarks where Gigabyte GA-EX58-UD4P falls significantly behind. However, we are not surprised to see it happen this way. We have already explained why Asus P6T does so well in our earlier review. Unlike Gigabyte mainboards, Asus increases the processor clock multiplier to 22 much more frequently. Due to this peculiarity of “Turbo Boost” technology implementation the board performs very well in synthetic Everest benchmarks. Unfortunately, this “advantage” has no effect on any other applications.

Now let’s check out the mainboards performance when the CPU is overclocked to 3.8GHz. Here most BIOS settings are adjusted manually to identical values. As a result, the boards perform almost identically, as we have expected:

The results of these comparisons show that Gigabyte GA-EX58-UD4P mainboard is exceptionally fast. It will outperform many competitor solutions, even those from the higher price segment. And in order to outperform Asus P6T, Gigabyte GA-EX58-UD4P may resort to its ability to overclock processors to 3.95GHz, which Asus couldn’t achieve. However, strange as it might seem, there is no convincing victory here for some reason. We only see a slight advantage of Gigabyte mainboard in a few applications, while in others it even falls seriously behind.

In fact, there is nothing strange about these results, if we recall that we had to sacrifice the memory frequency in order to push the CPU speed to 3.95GHz. The difference will be more illustrative in the next table that compares the performance of Gigabyte GA-EX58-UD4P mainboard with the CPU overclocked to 3.8GHz and to 3.95GHz. The CPU speed is 150MHz higher that is why we see about 3% performance boost in applications sensitive to computational capacity, such as Cinebench 10, Fritz Chess Benchmark, 3DMark Vantage CPU tests. Lower memory frequency explains a serious lag in Everest benchmarks. However, synthetics tests were not the only ones to suffer. We see the board slowing down in games, namely Custom PC Bench 2007 Multitasking Test that uses 7-Zip for data compression, which is sensitive to memory speed.

The obtained results once again prove a known fact that size (overclocking in this case) doesn’t always matter. Maximum overclocking results may not always ensure performance superiority. Sometimes, optimal parameters choice pays back better.