I'm absolutely aware how GPU boost works. I've overclocked several keplar, maxwell and pascal cards in the last couple of years and I also edited several boost clock states directly in bios. This doesn't make me an OC guru, but it gives me enough experience to know that you're running an overclocked card in the first place and to know what 3d mark/TimeSpy scores to expect.
You're also ignoring an important factor that I tried to explain to you now on two occasion: We don't know how fast the RX Vega is going to boost, the 1200 MHz is probably the standard non boost clock speed. Because just in case you aren't aware, AMD is also using variable clock states aka "GPU boost".
If you want to compare two different gpu architectures you have to either max both out or let them run at amd/nvidia specifications. Because any other way you can make things either look worse or better than they really are.
Quick example out of my 3d mark database.
GTX 980 highly overclocked: ~5000
GTX 1070 overclocked: ~6300
GTX 1080 running at nVidia standard clock: ~ 6700
The 1080 looks like a failure here and the 980 and 1070 like very good cards. But if you max out the 1080:
GTX 1080 highly overclocked: 8150
It's another story.
But you'll probably just ignore this, again.