A massive 6% behind the 980ti now, shocking!
Your math is wrong. (54-48)/48 * 100 = 12.5% perf/watt advantage for 980 Ti in 4K...pretty much the best case scenario for Fury X because it uses expensive HBM
Anyway, you can now get Fury X class performance from a much more power frugal 1070, which has been out for a full six months now. AMD has basically chosen to sit out of a product cycle.
I didn't bother to calculate the exact %, my point was more to show that your perf/watt comparison was not relavent anymore, as you used the oldest chart you could from the original Fury X review.
I linked one that's much more recent, with further driver improvements meaning it's only 12.5% (your math) behind the 980ti. Not the almost 30% that you first posted.
I notice you ignored my other post - why did you change subject from FuryX's and TitanX's power consumption to the 980ti? Hmmm?
the 1070 that even you, somebody who openly admits to preferring AMD, chose to buy.
if Vega doesn't definitively beat Nvidia, I don't know if AMD survives with high end gamers.
Fair points completely. So for me the VRAM is the major concern with Fury X. That and just the unbalanced nature of the card. With the hints that Polaris may actually have a lot more potential behind it, I'm interested in learning exactly what Vega is bringing. Either way, I just cant touch Fiji. It just performs too erratically.
No matter how low in price Fury X drops, GTX 1070 drops in price, and includes Watch Dogs 2. I imagine Nvidia will hit record profits again until Vega launches and if Vega doesn't definitively beat Nvidia, I don't know if AMD survives with high end gamers.
I don't see how Vega can definitively beat Nvidia. Maybe if Vega was released ~September... However, even if Vega is released @ Q1 it'll be too little too late even if it outright beats the 1080. Similarly to the Fury X - by the time it's released, they'll need a 1080ti/Titan competitor.
I'm not sure how much Nvidia even cares about AMD anymore. Unless AMD can bring a Titan XP competitor, they wont be very relevant, and I don't think such a card is really on the table.
AMD narrowing the performance/watt gap long after the cards in question are irrelevant like somebody who turns in his homework long after the class is over with and still manages to get fewer questions right than the student who turned it in on time.
Fury X provided worse perf/watt than a bog standard 989 Ti at launch, still provides worse perf/watt even in the best case, got wrecked in perf/watt and just straight up raw performance by aftermarket 980 Tis, and is now forced to do battle with the 1070 that even you, somebody who openly admits to preferring AMD, chose to buy.
Why are you so vehemently defending Fury X when you wouldn't even lay down your own cash for it?
So much outdated info and brainwashed comparisons...
AMD gets better and better with time, everyone knows that. Sweclockers, for example, uses May's drivers in that comparison for the Fury X. Here it is how it looks like these days:
With October drivers, Fury X performs 8% more in 4K and 2% more in 1440p, while being just 7.8% less efficient in 4K and 15% in 1440p than a stock 980ti.
And to get the hint of the efficiency of Vega, I will leave the updated potential of Polaris. Not the one we saw with the crappy first batches of the reference model.
We seen how amd totally owns with the 480 and Vega will be built even better and totally owns anything nvidia has.
its a reason amd radeon is the most used card out there.
You posted here that the Fury X "guzzles power" - when I pointed out to you that the Titan X and XP both us similar amounts of power, you start a rant about another subject, the 980ti, which wasn't even mentioned.
With a more mature 14nm process we are going to see Vega really push the performance per watt, but more importantly we are going to see massive wins in DX12 for AMD, even greater than what we have now. Hopefully DX12 becomes the primary api for engine and game development, not just an afterthought to be tacked on.
DX12 games can be so much better performing if developers just spent the time actually optimizing their games.
I can't be the only person who has bought MULTIPLE generations of AMD cards in a row who is TIRED of being told that newer games supporting newer APIs(or whatever the line of reasoning is) will be faster on your card. Then, have it end up being that same split of games that favor Nvidia and games that favor AMD....
I'm not denying that so far amd gpus have aged well with time, but by the time that time comes, I really don't care. It's a nice thing to care about if you're buying in the lower end brackets, but hearing the argument for a high end GPU.... makes me sad. We know how bad of a value proposition buying a high end GPU is to hold for a long time, so lets talk about what matters. The fact that an aging strategy won't work in the high end market. AMD needs decent performance or price from the getgo. It can NOT offer less than Nvidia performance for Nvidia price like it did last generation.
the 270x for example was 10x times better than its counterpart at 660. 660 and 660ti were overpriced turds, while 270x and 280x were much better alternatives depending on your wallet.
AMD did not offer slower cards in previous generation. 380, 390, 390X, Fury, had better or at least the same perf/price ratio compared to GTX 900 models. Only Fury X wasn't that good. Or was too expensive, depends how you look at it.
GTX 970 was probably the best selling card in recent years, and yet in almost all AAA titles from this year, it is slower than the R9 390. It looks to me 390 was smarter choice