As you can see, there's barely one percent of difference to report. Perhaps the only oddity is that, despite having the lowest BCLK, Gigabyte's board comes out a little ahead on the CPU benchmark.
Once again, there are insignificant differences in RAM bandwidth among these boards.
Nothing to see here, folks—move along.
The Heaven preset was my brief attempt to find settings that looked better than the defaults, without eating performance. And it looks like I succeeded, if only slightly. We still haven't found any real performance differences, though. That's finally about to change though; after all, I did promise some anomalies.
You can decide if it matters that Gigabyte's board slurps more power under a CPU load, running a little hotter in the process. It has the lowest BCLK, so I went back to the features to look for a cause. Is the LED underglow causing our anomalous results? Nope, that made less than a 1W difference.
Almost as an afterthought, I decided to take a brief look at network throughput. And since I don't have a NAS, I used the ubiquitous speedtest.net. I consider this an imprecise test in my current environment, which does not isolate how much bandwidth my neighbors might be using. Still, it might tell us something.
Gigabyte's network interfaces prove to be much faster than ASRock's offerings. Its wireless adapter even stomped on the Powerline speeds. Could that speedy wireless adapter also explain the added power consumption? I unbolted it from the board and re-checked. Nope. That was only a 2W savings.
Each of these boards is able to overclock the Pentium G3258. I hoped to get 4.2GHz, and did. But that overclock was not stable under Prime95 on any of the three platforms. I bracketed a colleague's settings and decided that I lost the silicon lottery this time. All of the boards were stable at 4GHz.
One by One
So, was anything else notable about each of these boards?