We usually benchmark mainboards in two test modes: at the default settings and with the CPU and memory overclocked. The first mode is interesting because it shows how well the mainboards work with their default settings. It is a known fact that most users do not fine-tune their systems, they simply choose the optimal BIOS settings and do nothing else. However, most of the LGA1150 mainboards we’ve tested so far call for some tuning in order to put them under the same conditions. We had to publish a long list of our adjustments in their settings, so the default test mode nearly lost its meaning. Instead of benchmarking the mainboards at their default settings, we would show the nearly identical results of our correction.
In the new series of reviews of LGA1150 mainboards we want to make our default test mode more informative. So we don't change or correct anything. What settings a mainboard prefers to use by default are used in our default test mode, even if they diverge from standard settings. It must be noted that in this case it is bad when a mainboard is slower than others, but it is also bad when it is much faster, meaning that it just doesn't use standard settings by default. It is desirable to get average results since we know that similar mainboards are going to deliver similar performance under the same conditions. We even thought about giving up showing the best results in the diagrams, but eventually decided to sort the results out in the order of descending performance. The results of the ASUS Gryphon Z87 are colored differently.
We run the CPU test of the 3D rendering suite Cinebench 15 five times and calculate the average result.
We have been using Fritz Chess Benchmark for a long time already and it proves very illustrative. It generates repeatable results and its performance scales perfectly depending on the number of computing threads.
x264 FHD Benchmark v1.0.1 (64-bit) helps us test video transcoding performance. The original version of the benchmark with the version r2106 coder could make use of AVX instructions but we use version r2334 to enable the new AVX2 instruction set available on Haswell-based CPUs. The results are the average of five runs of the benchmark.
We benchmark performance in Adobe Photoshop CC using our custom test that is based on the Retouch Artists Photoshop Speed Test and consists of typical processing of four 24-megapixel images captured with a digital camera.
The mainboards’ performance in cryptographic tasks is measured with the built-in benchmark of the popular TrueCrypt utility that uses triple AES-Twofish-Serpent encryption with a 500MB buffer. Besides optimizations for multi-core CPUs, it supports the AES instructions.
Metro: Last Light is a very beautiful video game but its frame rate depends heavily on the graphics card. So we had to use the Medium Quality settings to maintain playability at a screen resolution of 1920x1080 pixels. The diagram shows the averaged results of five runs of the integrated benchmark.
F1 2013 is less demanding on the graphics subsystem than the previous game. At 1920x1080 pixels we chose the highest settings by enabling Ultra High Quality and all image-enhancing options. The diagram shows the averaged results of five runs of the integrated benchmark.
The ASUS Maximus VI Hero is noticeably faster than its opponents in most of our tests, which is a sure sign of nonstandard settings. We know from our review of that mainboard that is automatically steps the CPU clock rate up by 200 MHz at multithreaded loads. It must be noted that we can get the same advantage on other mainboards by tweaking their Intel Turbo Boost settings. Gigabyte's mainboards can even offer more benefits in certain scenarios with their K OC option. It is easy to enable this if you want to, but it is not easy to disable such acceleration on the ROG series mainboard, so we are inclined to view it as a downside.
As for the ASUS Gryphon Z87, the short-term CPU frequency drops to the base level do not affect its performance much. The mainboard is comparable to similar products from other brands at typical loads when they all work at standard settings.