That test used a 6-core CPU and lacks Cinebench.
The effect of quad-channel depends on the workload and how many cores there are. Memory bandwidth requirements scale with the number & speed of the cores, in a workload with sufficient concurrency (such as Cinebench, which exhibits good multi-core scaling). Beyond concurrency, there's another workload-dependent variable, which is how memory-intensive it is. I think Cinebench might be one of the more memory-intensive tasks, as many of the others with good concurrency (e.g. WinRAR), can operate largely out of L2/L3 cache.
It's one thing to say that gamers needn't worry about quad-channel vs. dual, but it's another thing to say that the dual-channel config isn't hurting the Intel system in this benchmark. I think the latter claim is probably wrong.
Dozens of people must be doing this test, at this very moment. Let's see if we can find any results, and then we can put the debate to rest.