UPDATE:

Geforce GTX 560 Ti added with a rating of 129 Voodoopower

Radeon HD 6950 1GB also added with a rating of 135 Voodoopower

Let me know what you guys think!

Note: The improvements of GTX 570 with recent WHQL Forceware 266.58 drivers are not yet verified as to accurately change the ratings yet, above GTX 480. The same goes for HD 6970 and Catalyst 11.1 drivers.

What makes you place the Radeon above the GeForce overall?

The 6950 over the 560 Ti, you mean? Well, your review along with others, right?

EDIT: Upped the 560 Ti to 130 Vpower. Some reviews don't show it being much more than 20% faster than the 460 1GB.. I guess some games just don't scale very well, or the bandwidth/ROP's is starting to hold it back a bit once the full 384sp was unleashed. It only has a 11% increase in bandwidth over the 460 1GB, and the same 32 ROP's and 512KB L2 cache. But then again, it's more likely a scaling issue with some games like Call of Duty Black Ops, etc.. Also, the increase in TMU's from 56 to 64 does not seem to do it any benefit really, since it actually has more texturing power (64 TMU's @ 823MHz) than GTX 580's 64 TMU's while showing nothing for it.

I noticed that in some games that you reviewed here on ABT, there are substantial improvements with the newer drivers where the 570 is like 10-15% ahead of the 480 (like in CoD4 for example). Are there also big gains with the 6xxx series with 11.1 drivers? (I went ahead and gave the 570 one more point, tying it with a 6970, but will go over it again with final 11.1's).


Anyways, here are other reviews on the 560 Ti:
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/40119-nvidia-geforce-gtx-560-ti-1gb-review.html
http://www.xbitlabs.com/articles/video/display/geforce-gtx-560-ti.html
http://www.techspot.com/review/359-nvidia-geforce-gtx-560ti/
http://www.guru3d.com/article/geforce-gtx-560-ti-review/
http://www.anandtech.com/show/4135/nvidias-geforce-gtx-560-ti-upsetting-the-250-market
http://www.bit-tech.net/hardware/graphics/2011/01/25/nvidia-geforce-gtx-560-ti-1gb-review/1
http://www.hardocp.com/article/2011/01/25/galaxy_geforce_gtx_560_ti_gc_video_card_review/
http://www.computerbase.de/artikel/grafikkarten/2011/test-nvidia-geforce-gtx-560-ti-sli/
http://www.hardwareheaven.com/reviews.php?reviewid=1103
http://www.pcper.com/article.php?aid=1067
http://www.hexus.net/content/item.php?item=28597
http://lab501.ro/placi-video/gigabyte-gtx-560-ti-superoverclock-studiu-de-overclocking/8
http://www.legitreviews.com/article/1533/1/
http://www.tomshardware.com/reviews/nvidia-geforce-gtx-560-ti-gf114,2845.html
http://www.hardware.fr/articles/818-1/dossier-geforce-gtx-560-ti-contre-radeon-hd-6950-1-go.html
http://techreport.com/articles.x/20293
http://www.ocaholic.ch/xoops/html/modules/smartsection/item.php?itemid=453&sel_lang=english
http://benchmarkreviews.com/index.php?option=com_content&task=view&id=683&Itemid=72
http://www.bjorn3d.com/articles/NVIDIA_GTX_560_Ti_GF_114/1995.html
http://www.hardwareheaven.com/reviews/1104/pg1/asus-geforce-gtx-560-ti-directcu-ii-top-review-and-sli-performance-introduction.html
http://www.hardwareheaven.com/reviews/1103/pg1/palit-nvidia-geforce-gtx-560-ti-graphics-card-review-introduction.html
http://www.hardwaresecrets.com/article/NVIDIA-GeForce-GTX-560-Ti-Video-Card-Review/1180
http://www.hardwaresecrets.com/article/MSI-N560GTX-Ti-Twin-Frozr-II-OC-Video-Card-Review/1181
http://www.hitechlegion.com/reviews/video-cards/8121-nvidia-geforce-gtx-560ti-video-card
http://hothardware.com/Reviews/NVIDIA-GeForce-GTX-560-Ti-Debut-MSI/
http://ht4u.net/reviews/2011/nvidia_geforce_gtx_560_ti_msi_n560ti_twin_frozr_2/
http://www.insidehw.com/Reviews/Graphics-cards/GeForce-GTX-560-Titanium-is-Back.html
http://lanoc.org/review/hardware/video-cards/4757-nvidia-gtx-560-ti
http://www.hardwareheaven.com/reviews/1105/pg1/xfx-6870-black-edition-dual-fan-vs-inno3d-gtx-560-ti-oc-vs-gtx-570-vs-6950-introduction.html
http://www.neoseeker.com/Articles/Hardware/Reviews/NVIDIA_GTX_560_Ti/
http://www.ninjalane.com/reviews/video/nv_gtx560ti
http://www.overclock3d.net/reviews/gpu_displays/nvidia_gtx560_ti_review/1
http://www.overclockersclub.com/reviews/msi_gtx560_twinfrozr_2/
http://www.overclockersclub.com/reviews/nvidia_asus_gtx560ti/
http://www.pureoverclock.com/article1200.html
http://www.pureoverclock.com/article1202.html
http://tbreak.com/tech/2011/01/msi-n560gtx-palit-gtx560-zotac-gtx-560-reviewed/
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/40165-amd-radeon-hd-6950-1gb-review.html

It's literally taking me all day reading the reviews and factoring in the numbers. It's a mess.

BFG10K, help!! LOL..


BTW, Techreport has a review on the 1GHz version from Gigabyte. http://techreport.com/articles.x/20293/14 The power consumption is like 2% higher than the 570, while the performance is about 4-5% lower. It still destroys the 6950, though. At $274, it's a pretty good deal, considering the 570's $349 price. As long as it's not noisy, and actually under a cool temperature of 63C, wow!

All right, I had to type in all the numbers from the graphs that Apoppin did.
Here:


So, using your 29 games (most were using the highest resolution, as long as it were playable, if not, then I selected the lower resolution for that game), the 6950 is 6.04% faster than the 560 Ti.

If I leave HAWX 2 out (a unusually optimized game in NV's favor), it changes to almost exactly 7%.

Hey, do you mind sharing your whole excel spreadsheet with all of this data, so that I could just sum the numbers up for other cards like the 6870 and 6970, etc.. saving me a lot of trouble?

(BTW, your Far Cry 2 only shows the 1920x1200 chart 2 times.)

That IS an awful lot of work


.. sure, i can send you my Excel Graphs
- i'll clean it up a bit ..

Maybe you could make the Performance Summary Chart



Why leave out HAWX 2? You should also leave out Wolfenstein ... FC2 seems to favor nV HW like F1 2010 favors Radeons

LOL, just curious about the excel data.. once the data is in there, it's easy to do a sum of each column. Only takes 1 second to do that per column, so it'd save me LOTS of time.

About the performance summary chart, I'd try if you want me to, but I don't think I know how to make it look like the one you did on your ABT reviews. Just raw summary data?

i was kidding about you doing my charts


let's see if i can get the charts off to you tonight .. in between final proofing my article

i sent you the charts. i hope it helps

I believe it's more accurate to first find the % difference between two cards per game. Then average that % difference.

Summing up the framerates actually gives more weight (as in, weighted average) to games which produce higher framerates (higher numbers). Find the % difference per game ensure each game as the same weight in the final average.

But really it's not that big of a difference, simply because when we test most games they are getting roughly the same framerate anyway (30-90). But the more accurate approach is what I have already stated. You can plug random numbers into excel yourself and see the small difference. Like plug in a bunch of numbers in the 30-50 range for two columns. Then, let's say, one card gets 100 and another get 150, and you'll see the difference I'm talking about and how that's reflected in both methods.


Actually here's a googlesheet illustrating the point:
https://spreadsheets.google.com/ccc?key ... y=CKTBkIoM

Since the 100-150 results have a difference of 50 in t3, that's having the same weight in the "quick sum" method as five tests that only have a difference of 10 fps. But really, five different tests should each have equal weight and be reflected in the final calculation. Since there are more times card 2 is 50% faster in the second chart, that should be reflected in the final average, which is if you give each game the same weight but not if you just sum everything.