The motherboard's QVL-certified memory will often be the safest bet and least hassle. You can sometimes do better but ASUS has already tested almost every kit and part number available.
I've discovered that G.Skill sometimes changes their kits without changing the part number. I still have 4x8GB DDR4-3000 kits (from early 2014) which use Hynix silicon and generally outperform 4x8GB DDR4-3000 kits (from mid 2015 onwards) which have exactly the same Ripjaws4 part number but use Samsung silicon. I also have newer same-kit-same-part-number Samsung parts which generally outperform older Hynix parts. Just trying to say that it's sometimes hard to know all the specifics without digging through tons of reviews/etc. I wouldn't be surprised if Corsair or other-branded memory sometimes does the same thing (in fact, I'd be surprised if they didn't).
Arne hit the nail on the head - the biggest bottleneck for upper-end RAM performance is usually the CPU part. If you've got a strong piece of silicon then you can run faster DDR4 ... if you've got a lemon then you can't. X99 mobos can usually run 4 DIMMs with an "average" CPU at up to -3000(ish) or -3300(ish) at most anyhow, not really worth paying the premium for faster-rated stuff unless you plan to re-use it in other platforms.
"All opinions are not equal. Some are a very great deal more robust, sophisticated and well supported in logic and argument than others." - Douglas Adams
[/Korth]
