GDDR6X hasn't really shown itself to be the huge advantage that nVidia touted it to be. If it were, nVidia would win at everything, but it doesn't. This just proves that GDDR6X is barely more of an advantage than HBM (if at all). It's as much of a marketing buzzword as HBM was for AMD back in the day. At the end of the day, the same question and answers apply to GDDR6X as they did to HBM and HBM2:
- Q1: Is it faster than standard GDDR?
- A1: Hell yeah!
- Q2: Does it matter?
- A2: Hell no!
To date, I've never seen a clear example of faster VRAM making a significant impact. ATi was the first to introduce GDDR5 but its speed advantage over the GDDR4 that nVidia was using didn't translate into better gaming performance.
The HBM that AMD used on the R9 Fury series was an order of magnitude faster than the GDDR5 used on the GTX 980 and GTX 980 Ti. The GTX 980 and 980 Ti used a 384-bit VRAM bus which is quite a respectable bus width for a video card. However, comparing it to the width of the HBM VRAM bus in the R9 Fury and Fury-X was like comparing a back alley to an expressway. This is because HBM uses a staggering 4096-bit bus. Regardless, the Fury series still found itself sandwiched between the GTX 980 and GTX 980 Ti in gaming performance.
Later on, along came the much-anticipated Radeon VII with its massive 16GB of HBM2. Now, HBM2 uses a 2048-bit VRAM bus but that's still gigantic. I don't need to tell you what a flop the Radeon VII was in gaming, easily bested by the GTX 1080 Ti.
The more I look at these cards with their "fancy" VRAM types, the more I think that gaming performance comes from having a very fast GPU and VRAM that is fast enough not to bottleneck the card. In that regard, GDDR6 is plenty fast.