They will mix fine and don't worry about the whole dual vs single issue, it's vastly overrated. Real world benchmarks will show the difference is so small you wouldn't notice the difference. As stated
here they show dual channel only offers a 2%-4% difference, meaning even long tasks that take say 10secs, it's not even 1/4th of a second faster. And in your most demanding games you might get 1 fps more at best. In your older games where you'd gain maybe 5fps, you're already well above 60fps so it won't matter.
Personally, I'd put the largest chip in first (which is usually the newest anyways) as the older chip will go bad first, and if it does, you're less likely to use it, meaning less memory errors as it fails. But I'm splitting hairs here.