They show cased Polaris almost 7 months before it's official launch. They were "modest" with their showcase and through out the product showcase outside of some boastful performance / watt claims, they kept relatively modest with their expectations.
But the hype monster and AMD fans WANTING a win, hell take a stroll through the CPU subsection, basically turned Polaris 10 in a Fury X replacement. Even when leaks started to funnel out, it was unbelieveable by some that Polaris 10 would just meet 390 performance and in some occasions 390X.
Polaris was disappointing technologically (for whatever reasons, arch related or fab related), with perf/watt and perf way worse than what people expected. If this is what's in store for Vega 10, I just don't think AMD has anything to market if they wanted to.
If Polaris was a major disappointment to anyone performance wise, that is only because of the above mentioned hype machine getting out of control and setting expectations way too high. AMD has been absolutely consistent about the 480 offering 390 level performance at reduced power consumption and that's exactly what we got. I honestly thought they would get more praise for being upfront about what Polaris is since people are always keen to jump on a company for over-promising and under-delivering.
Granted, there's things to criticise about Polaris, mainly the perf/w being somewhat disappointing, though I still don't think it's quite as bad as some people have made out. Compared to previous gen AMD cards it's a substantial improvement and I expect there's still room for more efficiency with better drivers/undervolting/a new revision of the cards. Polaris' perf/w may not look so hot (ha!) against Pascal, but at that point I feel I have to point out the obvious, that Nvidia have practically limitless R+D resources against cash-strapped AMD, they've focused heavily on efficiency compared to AMD cards having more raw computing power, and by all accounts their 16nm process is a better one than whatever shenanigans AMD are stuck with at GloFo. Given all that I think it's at least a bit unfair to compare Polaris so unfavourably to Pascal.
I tend to feel that while AMD are a long way from being a flawless company, even when they execute relatively well they still can't win. Look at the stock situation with Polaris. They could have launched along with Pascal, I'm sure of it, but the availability would have been even more dismal. Result? 'Paper launch, AMD fail'. They could have waited another month for stock to build through July, result 'Late launch, NV beat them to the punch, AMD fail'. Suggesting that they should have had more cards ready, okay, review again the disparity in resources compared to NV, now is anyone going to seriously suggest AMD have been sitting around thinking 'well we could launch these millions of cards we have gathering dust, but let's not'? (reading the AMD reddit recently apparently there are people who believe this; 'AMD should make more cards!' Yeah, thanks, I'll pass that on to Lisa Su whenever I get the chance. Don't even get me started on how vilified miners have become on there).
Vega then? I can already see the pattern repeating. I don't think it's going to be a bad card, in fact I think in terms of pure performance per dollar it's going to be highly competitive with NV, but I can say for a fact right now it
is going to disappoint people aboard the hype train expecting some miracle. Rinse and repeat for Zen