I want to counter what you are saying (because Nvidia gives me bad mojo here) hahaah, but I think the easiest way to look at this would be to see their profit margins per/year vs devices/aka GPUs sold. If you saw a large increase in profits (which has happened) with a similar amount of GPUs sold/similar amount of types of gpus sold, then we would know for certain which narrative holds the most water. I am sadly going to take a nap now because I am tired, so if anyone would bring up that data I would be grateful!

I don't think AMDs cards will be that much cheaper now tbh, as they aren't skimping on the coolers anymore, they are putting those Asetek 295x2 hybrid coolers on em.

You say dodgy choice starting with G80 then you link to a benchmark @ 1280 res, lol.

It was actually RELEVANT at the time. http://tpucdn.com/reviews/Zotac/GeForce_8800_GT/images/3dmark06_2048_1536.gif Congratulations. 3 Million pixels per frame mean that now I'm apparently not dodgy IYO.

It was actually RELEVANT at the time. Congratulations. 3 Billion pixels per frame mean that now I'm apparently not dodgy IYO.

Idk I had a 1080p monitor since 04. And I started at G80 because the wiki entries for anything prior don't have nice charts with release date timings and I'm lazy MUST MEAN IM "DODEGY, IYO". The 7800GTX was launched at $600 though in June 2005. The 6800 Ultra launched in early 2004 for $550. The performance difference between a 6800Ultra and 7800 was 50%. You linked the years but you didn't link the gaps between them. The FX series didn't have an increase in performance. 6 - 7 was roughly a 50% increase. Previous ones were at the infancy of computer graphics. The phone CPU market is going through the same thing now, you aren't seeing doubling of benchmarking speeds year after year. Regardless it doesn't even matter. I said this "It's easy to look at Nvidia and just say "well their greedy" but there are a lot of manufacturing and engineering constraints going on behind the scenes." and I stick by it. People keep trying to simplify it down and look at basic historical data for answers but the bottom line is the economy, technology and nvidia as a company are in a very difference place now then they were in 2006. You can't just simply turn everything into an economics 101 graph of "profit vs unit sold" or "performance vs price" then make blanketed statements about the state of the industry, company or even a specific product. Looking at it from a performance perspective is even worse. The cost of developing a new architecture is going to increase exponentially as chip design becomes further complicated. Couple that with diminishing returns from die shrinks, price changes, shifting markets, every other thing I've been posting about, it's obviously going to have a significant impact in development of performance. If people were just saying "well it sucks that all these changes led to slower then usual performance on modern GPU's" then I wouldn't mind. But they aren't, they are attributing it to "greed" as if Jen-Hsun is personally pocketing all the profit.

$600 will be BAD due to GM204 = full mid-end = no ultra memory bandwidth bus! :thumbdown

$600 will be BAD due to GM204 = full mid-end = no ultra memory bandwidth bus! :thumbdown

Will absolutely smoke a 580. The 580 bandwidth is lower than a 770 980 970 960. Clock speed of 7ghz+ is faster than a 580 384bit. My 680s were about 40% faster than my 580 sli both 580s and 680s both clocked to their max.

Will absolutely smoke a 580. The 580 bandwidth is lower than a 770 980 970 960. Clock speed of 7ghz+ is faster than a 580 384bit. My 680s were about 40% faster than my 580 sli both 580s and 680s both clocked to their max.

Not to mention things like L2 cache increases and better compression. Even if 580 had the same bandwidth number as a Maxwell card the Maxwell card would still have better bandwidth performance.

I've been completely over the whole speculation thing. The specs posted this morning are better than I had anticipated, the $599 cost is still speculation, and even if it is true I don't think it's too much personally. Of course I would like it to be $499. Rumor has it that NO benchmarks for the 980 have been posted and what we have seen are actually the benchmarks for the 960 and 970, and that the 980 is actually going to be a 15-20% increase in performance over the 780ti. If all of that turns out to be true I don't know what anyone is crying about. A $499-599 780ti killer that uses much less power and with more ram.. I think too many people are looking at individual specs (the ram bus) and going nuts over it.. wait to get the whole picture.. that 384bit bus just might not be necessary. Is it really too crazy to think that Nvidia themselves have been trolling us with false-ish info so that the actual launch event will be exciting? Can't wait for the actual benchmarks and per game FPS performance charts to start rolling in this weekend. Any which way you slice it, this is going to be a killer upgrade from my 580's 😉

Will absolutely smoke a 580. The 580 bandwidth is lower than a 770 980 970 960. Clock speed of 7ghz+ is faster than a 580 384bit. My 680s were about 40% faster than my 580 sli both 580s and 680s both clocked to their max.

Ok but GTX 980 can't be faster memory bandwidth than GTX 780 series. :thumbdown 😕 I'm much deserved the HIGH-END (= each next-gen card is ALWAYS faster memory bandw and eg: GKxx0, GMxx0, GVxx0 (Volta chip)...). 😉 EDIT: GDDR5 will continue "badly" since 2010... Still no XDR?! 😕 🙁

Ok but GTX 980 can't be faster memory bandwidth than GTX 780 series. :thumbdown 😕 I'm much deserved the HIGH-END (= each next-gen card is ALWAYS faster memory bandw and eg: GKxx0, GMxx0, GVxx0 (Volta chip)...). 😉 EDIT: GDDR5 will continue "badly" since 2010... Still no XDR?! 😕 🙁

ok fine it's gonna fail, you won't have to upgrade till 2015 at least, or whenever you want. no one is forcing you to lol

Not to mention things like L2 cache increases and better compression. Even if 580 had the same bandwidth number as a Maxwell card the Maxwell card would still have better bandwidth performance.

My 680s at 7200 had quite a bit more bandwidth than a 580, though memory oc on the 580 didnt help much at all. I ran the memory at 4410mhz for 2 years

WOW! WCCFTech is confirming the price of the 970 to be $299 and custom cooled variants up to $349! Is it at all possible that Nvidia has been listening? With that price it seems the $499 speculation for the 980 might possibly be high.

Didn`t they also confirm the $299 680 or was that another site?

I think that was fudzilla. Of course the 680 was not $299 when it released lol.

People are still confused about that 680 if you ask me...

I think that was fudzilla. Of course the 680 was not $299 when it released lol.

I paid $559.99 each for 2 680 ftw+, I can confirm it was not $299 🤓

it costs nvidia less now to make the cards but they increase the price:stewpid:

NO THANX:stewpid: I feel the performance improvement is so marginal than to its price point improvement, every time! wish its otherway around:puke2:

Ok but GTX 980 can't be faster memory bandwidth than GTX 780 series. :thumbdown 😕 I'm much deserved the HIGH-END (= each next-gen card is ALWAYS faster memory bandw and eg: GKxx0, GMxx0, GVxx0 (Volta chip)...). 😉 EDIT: GDDR5 will continue "badly" since 2010... Still no XDR?! 😕 🙁

There will never be card with Rambus' memory. Next step is stacked ram and I'm quite sure that the next gen will use stacked ram. Memory bandwidth on paper is not everything. This will be ****loads faster than GTX 580 and on top of that according to leaked tests this is computing master (much faster in direct compute).

.. on top of that according to leaked tests this is computing master (much faster in direct compute).

oh hell think i'm starting to get tempted ..wait yes i'm totally tempted now. http://i.imgur.com/TgVK0Kw.png