Well... I have such an excruciatingly bad feeling of déjà vu, that I wanna scream. If the leak and the performance is legit and big navi would be anywhere between 20-30% faster than the 2080ti (because the fastest 2080tis are heavily overclocked on that leaderboard), then the only thing it would achieve is to give NVIDIA the perfect opportunity to copy their 980ti moment (Big Navi is 20-30 faster than 2080ti at around 2080ti power consumption, but then comes the next NVIDIA flagship to be even faster and much more efficient) to an extent where RTG is ridiculed in the eyes of the enthusiast community. I mean, 980ti and Fury X were at least close in performance, but of course the 980ti used waaaaaaaay less power.
Remember, Pascal, for example, came on the 16nm node in a way that even the very last week before launch nobody really wanted to believe it's actually been taped out, and after launch everyone was convinced that it's only a paper launch and RISK production etc... Still, they shipped millions in a short time. NVIDIA is not Intel, I don't know why, but they seem to prevent leaks very efficiently (maybe everyone is afraid of Mr. Leather Jacket!!!
), and even when everything points to them not really having wafer allocation at TSCM, and that Samsung's HP 7nm EUV is in a dire shape regarding yields, all that just reassures this dreadful déjà vu I'm having.
But as always when I fear for AMD (they still need a lot of financial success to remain a competitor to the 2 players for more than just 3-4 years), I really hope I'm not right
Also, new consoles as RTG/RDNA's saving grace? We'll see.
The only other idea I have is to somehow make their GPUs 6-wide with SMT4 and maybe they can catch up to Apple's IPC. Oh wait....
Joking aside, I have no problem seeing big navi as an NVIDIA killer - if that means to have finally beaten a (by then) 18+months GPU on 7nm vs 12nm. How do they want to price such a monstrous 7nm GPU right for all this to actually mean anything?