Yes, they're charging a premium for lower power/very quiet operation. So? There's a non trivial market for that. I haven't noticed Noctua etc giving their fans/coolers away
Yes, but what a premium!
1080P
GTX960 = 66% = $200
GTX960 SLI = 100% = $400
R9 290X after-market = 103% =
$270 (
56% faster at 1080P than a 960 for only 35% more money)
GTX980 = 120% = $550 (
17% faster for at 1080P than a 290X for 2.04X more money)
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_960_SLI/23.html
You can now buy up to
60% more performance with dual 290Xs for the same price as a 980. GO PREMIUMS!
If power consumption of 2-3 light bulbs is so critical, why game on the PC and buy expensive videocards?
Wii U uses 35-37W of power total vs. 230-240W of power of a 960 PC and it's quieter than any 960. Clearly the Wii U is way better than a PC with a 960 as it uses 6.5X less power, runs quieter but no way does a PC with a 960 produces 6.5X better graphics.
We could then conclude erroneously that Wii U >>> Gaming PC with a GTX960. See how talking about perf/watt, power usage becomes a subjective discussion and leads to some crazy conclusions?
BTW, the 960 is just 9% faster than a 760 and the overall system power usage is just 30-40W less. You made it sound like a 960 level of performance/watt is some
revolution or something. If you took 2 gaming systems and put them side-by-side and one used 230W of power with 9% more performance and the other used 270W of power, you would never be able to tell while gaming in the room without a Kill-a-Watt and benchmarking.
Right now you can buy a cool and quiet R9 290 (You still can't accept that
such cards actually exist), and drop the clocks to 50%, set an FPS cap to limit your power usage. When you want ultra textures and higher performance, you can have that too. Can you do that on a 960? Nope, you have no such flexibility as you are stuck with lower quality textures and slower performance.
When next gen games come out, I guess you'll be fine to spend
another $200 for a 120W x60 Pascal that might match a 3 year old $250 R9 290?
It's not just about saving electricity today, but about paying a little extra so that you don't have to drop yet another $200-250 so soon as your videocard becomes way too slow, way too soon. By 2016, a $200 x60 Pascal might beat an R9 290 but it'll cost you another $200 but the R9 290 owner just spent $50 for similar performance and enjoyed all of that for almost 2 years!
Is faster simply because Kepler got a bit outdated. In older games GTX 960 will be as fast as GTX 760.
I think it's still valid to use the data from TR and TPU but their GPU review conclusions have to be taken with a huge grain of salt imo. Essentially the authors prioritize perf/watt above all other metrics, but TR is even worse since they ignore frame times, the very topic they bashed AMD on for 3 years after
NV provided them with the FCAT tools. Also, the editor of TR loves Borderlands games, which have PhysX. He's stated many times that it's among his favourite franchises to play. I don't see him as being an impartial gamer since he obviously cannot play without PhysX in his favourite game series. It doesn't take too long to connect the dots that your favourite game series runs better on NV so what cards is he more likely to recommend in his reviews?