And now let’s run the same exact tests in overclocked mode, when both - processor and memory – work at higher frequencies. Remember, that ECS P67H2-A2 and Gigabyte GA-P67A-UD4-B3 overclocked our processor only to 4.7 GHz, on Foxconn P67A-S we forced overclocking without the increase in the processor Vcore, so we had to stop at 4.5 GHz, while on other mainboards the CPU frequency was increased to 4.8 GHz. The system memory worked at 1600 MHz with 6-6-6-18-1T timings on all testing participants except Micro-Star P67A-GD80 (B3). Here the memory frequency was increased to 1867 MHz and the timings were set at 7-7-7-20-1T. The table below shows the differences in system settings for each testing participant very clearly:
This time the results of ECS P67H2-A2 should be compared against those of Gigabyte P67A-UD4-B3, because both these mainboards worked in very similar conditions and overclocked the processor to 4.7 GHz. Elitegroup mainboard performs very well, running almost as fast as Gigabyte board. Moreover, both these mainboards are just a tiny bit behind the other participants which managed to push the CPU clock to 4.8 GHz. ECS P67H2-A is among them, too, and its speed is the same as that of others except for the 3DMark 11 and games. Just like in the nominal mode, the performance in VGA-dependent tests is so low, that ECS P67H2-A loses not only to the products in its group, but also to all other mainboards including Foxconn, where the CPU overclocked only to 4.5 GHz. So, it turns out that Lucid Hydra chip can do not only good. As a result, it won’t be optimal to use a single graphics accelerator on mainboards like that, because of the overall slowing down. Frankly, the performance difference is not too dramatic, and doesn’t exceed 3%, but it is obviously visible and undeniable.