blazorthon :
Your "favorites" are extremely outdated and not relevant in the face of current drivers.
I went with the apples and apples thing .... I disregard sites (like the one most often referenced in support of this "new drivers" position) that test one card's new drivers against another card's two generation old drivers. Can't really understand the championing of wonderful new drivers when the reader didn't bother to check or intentionally ignored what driver was used for the competition on the new test. I also see little relevance in "optimizations" that sacrifice image quality in order to tweak benchmark scores.
Not saying that AMD is the only one guilty of this tactic, nVidia has done it's "optimizations" too over the years. Benchmarks become more useless month after month as
both card makers tweak their drivers to perform well in popular benchmarks..... but that's all you get .... improvements in benchmark performance. Do the same test for a different portion of the game and these new optimizations often don't change beans.
http://benchmarkreviews.com/index.php?option=com_content&task=view&id=12845&Itemid=47
http://news.softpedia.com/news/AMD-Accuses-Nvidia-of-Cheating-64728.shtml
Should we also be more concerned about what most reviewers are not testing ? ..... Should we care more about latency than whether one cards gets 79 versus 75 fps for the other ...... it's those "nasty spikes into the 50 and 60 millisecond range" that put the downer on the gaming experience, so this is a factor I will be looking to see reported on more in future testing. Start at the link below and follow thru the rest of the benchies....in most cases, they compete well ...90 - 95 % of the time....in the other 5 - 10%, we see moderate to severe lags
http://techreport.com/review/23981/radeon-hd-7950-vs-geforce-gtx-660-ti-revisited/3
Last round (GTX5xx / HD6xxx), it was easy ....nVidia above $200 and AMD below. This round, to my mind, since the AMD took the axe to its pricing structure, it's hard to make the wrong choice
on a cost per frame basis no matter which way you go. The user's I typically build for (speaking strictly of $200 cards and up) seem to prefer the nVidia cards about 2:1 over the corresponding AMD card (this ratio is echo'd by what is hitting steam servers and hence what people are buying ... see link below) but I don't have much invested in their choice so don't have an "emotional" involvement.
http://store.steampowered.com/hwsurvey/videocard/
GTX 670 has 2.11% of the DX11 market, up 0.17% in the last month
HD7970 has0.81% of the DX11 market, up 0.05% in the last month
HD 7950 has0.77% of the DX11 market, up 0.08% in the last month
The nVidia / AMD battle today is as competitive as I can remember. New driver arguments are only valid however when:
a) The tests use current drivers from both manufacturers
b) The tests don't limit their testing to a short list of games so as to produce a predetermined conclusion
c) The test settings are not selected to favor one side or the other.
There are many people from both the red and the green teams who are quite happy to ignore all the above if it produces their desired conclusion. Ya don't like my links (I chose them for the wide number of games and wide variation in settings), post your own. My advice to the potential purchaser however remains the same..... look at many reviews.....look beyond the surface into the reviews and:
1. Check both manufacturer's web sites to make sure that all drivers used in the test are current.
2. Ignore any site / review with < 10 games in their testing protocol ... the more, the better.
3. Give more weight to sites that use variety of settings and resolutions to paint a complete picture.
4. Look at the individual games in the tests and give greater weight to the ones you are likely to play.
5. Look at other factors besides average fps.
6. Don't accept statements in forum posts that don't have links supporting the position..... and evaluate the links based upon 1 - 5 above.
Again, there is no "wrong choice" in the market at this time......do ya research,
validate it and make your own choices.