as long as your sticking to 1080p anything 970 or higher outta carry you through this console cycle just fine
Ya, if you are willing to turn down settings, you can even use an HD7970 all the way until 2019 because PS4 is slower. A lot of people don't like jaggies in games and don't want to play at 30-45 fps. I am not saying that 60 fps minimums/averages is mandatory for everyone or that playing without AA is blasphemy but the point is IF a gamer wants those things, then even today 970 isn't enough for 1080P in some titles.
https://www.youtube.com/watch?v=53mWbxcWAw4
I am pretty sure that 970/290 will start to struggle in modern AAA PC games in 2017, and esp. in 2018 and 2019. I am not expecting to see PS5/XB2 until Fall 2019. 970 is 'only'
35-40% faster than the R9 280X. Even with overclocking, that's not a lot of headroom for next gen games. To put this into context, we had 280X level of performance all the way in early 2012, or almost
4 years ago. In comparison to today, 970 is just
40% faster but it has to survive another
4 full years? I mean it will but at very reduced settings.
By end of 2016, I wouldn't be surprised if it would be possible to purchase R9 390/970 level of performance for $160-170.
QUESTION
how many of you truly believe in future proofing?
Here is my take:
1. If on a budget, the best way to future-proof is to buy a card that's good enough for the games you play @ your preferred resolution, fps, IQ settings without overspending. Then, for next gen games, sell this card and get something better. A good example of that is say a $299 HD6950 unlocked to a 6970, without wasting $ on a $499 GTX580. Another great example is buying a GeForce 4 Ti 4200 and overclocking it to GeForce 4 Ti 4400 speeds, or $280 HD7950 OC vs. $450 GTX680 2GB. If you can get 85-95% of the performance for hundreds of dollars less, it's smarter to put aside $ towards the next gen upgrade.
2. Sometimes paying a little more can net big performance gains or crucially VRAM that matters. Some examples of that would be 8800GT 256MB vs. 8800GT 512MB or HD4870 512MB vs. HD4870 1GB. Another example applies very well to this generation: $600-650 R9 295X2/R9 290X CF vs. $550 GTX980 or $250-260 R9 290 vs. $200 GTX960. In some cases, it's cheaper to spend a little more because you'll get much more useful life out of the videocard. If such a scenario applies, "future-proofing" by paying slightly more can have big payoffs. But sometimes paying more is unlikely to have big pay-offs such as $270 R9 390 vs. $480 Fury. Think about how you can just put aside $210 towards a next gen card that is way faster than the Fury.
as in buying something faster than what you currently need in hope of not having to buy again in the near future.
why and why not?
To counter my previous 2 examples, there are cases where overspending on a new card slightly faster will not pay off. Obviously HD6970 vs. 6950 or paying $499 for a GTX580 when GTX480 is $299. Another example is GTX980 vs. 970 or HD7970Ghz vs. the standard HD7970 or R9 290X vs. R9 290, etc.
The strategy of buying cutting-edge graphics cards at launch could become costly quick.
About 2 years ago:
GTX780Ti SLI = $699 x2 = $1400
R9 290X CF = $550 x 2 = $1100
R9 295X2 = $1500
Just 2 years later, you can get basically the same level of performance for
$400. We also already saw GTX980TI drop to $530 over Black Friday. That means nearly R9 295X2 level of performance for
1/3 the price.
In that context, spending
a lot of $ to future-proof isn't worth it.
this setup pretty plays everything at stock ultra setting 1080p at enjoyable frame rate (beside crysis3 - which is not a primary game).:
Option 1: NV path
-> Kids rig = sell 1x 280X for $100, keep 2x 280X. There is going to be very little benefit to having a 3rd 280X for a 1080P gaming rig for your kids, so what's the point of keeping it?
It's hard to replace 2xR9 280X with something much better on a reasonable budget so I think it makes sense to just keep these 2.
-> Your rig = sell 4x290X for $200/ea = $800
$100+$800 = $900
Now buy 2x980Ti
Next time you upgrade, you can sell 2x280Xs in your kids rig, stick one of your 980Tis in there and then you can sell the 2nd 980Ti you have left and buy 2x flagship 16nm HBM2 cards.
Option 2: AMD path
Same thing as above, just replace "980Ti" with "Fury X"
Option 3: AMD + NV path
-> Sell 3x280X for $100/ea = $300
-> Transfer 2x290X to kids rig
-> Sell 2x290X for $200/each = $400
Buy 980Ti SLI / Fury X CF. This way your kids and you get an upgrade but if your kids are already happy with the performance of their system, Option 1 or 2 sounds better imo.
As far as 4xFury Xs go, I am not seeing the point of that. If you wanted 4 top-of-the-line cards for $650 ea., I think it makes the most sense to buy them as close as possible to launch date.
You could also just wait until 2016 and upgrade then. Chances are neither the 280Xs nor the 290X will fall much below $100 / $200 resale values. However, there should be cards at least 25-30% faster than the Fury X.