Real Results: Battlefield 3 Multiplayer 64/64p 6970 vs 7950 vs 7970 @ 1200/1700 + mor

2wiced

2[H]4U
Joined
Feb 14, 2005
Messages
3,367
*DISCLAIMER*: ****These tests are from 64 person multiplayer games. Depending on your play style your FPS may vary. For example, piloting a chopper takes a toll on frame rates. Being in action with explosions all the time also takes a toll on frame rates. In these tests I tried my best to stick to as much action as possible. That means explosions, explosions, explosions...and a couple of suicides lol. I hope that you enjoy the numbers below. Thanks again for reading!******

Hey guys.

So for a few days I have both cards. A reference 7970 and a reference 7950. Both can reach 1200 core/1700 memory (reference card can go higher about 1250-1275 core and 1775 ram).

Right now im on a stock intel cooler. So I cannot OC my 2500k too far. That is a limitation here. I am testing online multiplayer the best I can. So please understand that these results are not perfect.

2500k @ 4.2ghz 8gb DDR1600 128gb SSD Crucial M4

22 inch monitor 1680x1050

img07461n.jpg


Top to Bottom: 6950 unlocked shaders (950core1450mem), 7950 (1200core/1700mem), 7970 (1200core/1700mem)


7950 stock 800/1250 vs 7950 1250/1850

cb7950.jpg

gulf7950.png

More results

CASPIAN BORDER

Radeon 7950 @ 1250/1850 + Tessellation ON

Frames Time (ms) Min Max Avg
12455 158809 52 160 78.428


Radeon 7950 @ 800/1250 + Tessellation ON

Frames Time (ms) Min Max Avg
7960 152584 33 78 52.168




minimumfps.jpg

averagefps.png

tessoff.jpg


Below are results with tessellation ON CCC application preference
CASPIAN BORDER

Frames Time (ms) Min Max Avg
22084 283236 52 122 77.97


Frames Time (ms) Min Max Avg
16950 240757 51 122 72.403


Frames Time (ms) Min Max Avg
14335 332282 31 81 43.141


GULF OF OMAN

Frames Time (ms) Min Max Avg
38330 553289 36 144 69.277

Frames Time (ms) Min Max Avg
27093 419768 38 105 64.543

Frames Time (ms) Min Max Avg
12581 311160 23 70 40.433


KHARG ISLAND

Frames Time (ms) Min Max Avg
27811 407849 46 163 68.189


Frames Time (ms) Min Max Avg
30008 442840 47 116 67.763


Frames Time (ms) Min Max Avg
20794 432996 30 138 46.024


Alien Vs Predator D3D11 Benchmark

Number of frames: 13991
Average Frame Time: 7.5ms
Average FPS: 133.3


Number of frames: 13458
Average Frame Time: 7.8ms
Average FPS: 128.2

Number of frames: 8148
Average Frame Time: 12.9ms
Average FPS: 77.7


Below are results with Tessellation OFF via CCC
KHARG ISLAND
7970
Frames Time (ms) Min Max Avg
37204 426726 50 167 87.185


7950
52999 669900 50 171 79.115

GULF OF OMAN 64 player

7970
Frames Time (ms) Min Max Avg
99802 1333980 35 200 74.815

7950

Frames Time (ms) Min Max Avg
56610 759787 36 173 74.508

CASPIAN BORDER
7970
Frames Time (ms) Min Max Avg
74584 1009248 47 139 73.901


7950
Frames Time (ms) Min Max Avg
70338 895617 44 168 78.536





3DMARK11 For fun. lol.

GraphicsTest1 44.53 41.29
GraphicsTest2 56.62 52.65
GraphicsTest3 70.42 67.46
GraphicsTest4 41.67 39.98
PhysicsTest 25.07 24.96
CombinedTest 36.24 36.6 FPS
 
Last edited:
Good to see a real apples to apples both overclocked that high. what fan speeds did you end up having to run with each card? I'm they sounded like jet engines being reference cards haha.
 
Good to see a real apples to apples both overclocked that high. what fan speeds did you end up having to run with each card? I'm they sounded like jet engines being reference cards haha.

The 7970 is a reference card and I have a custom fan speed set through afterburner. I game with G35 headset so I dont hear the card when I game. If I take off my headset, then it sounds like a hair dryer lol.

The 7950 is the sapphire 7950 OC.

Im surprised my that so far they are pretty much identical in performance. The 7970 seems smoother in the game a little. But the minimums are the same. Gulf of oman played identical on each one.

Also. My GPU usage remained 99% on each card throughout gaming. So I know the CPU isnt bottlenecking either one of these cards.

Ill do some more BF3 in the next hour or hour and a half. The trouble is trying to find a 64 person server thats open lol. Right now im installing 3dmark11 for shits and giggles.
 
Can you compare the two cards in an Eyefinity set-up?? On a single monitor, the 7950 really shines. Great job comparing the 2 cards at higher clocks.
 
Can you compare the two cards in an Eyefinity set-up?? On a single monitor, the 7950 really shines. Great job comparing the 2 cards at higher clocks.

I dont have 3 monitors. Sorry.

My resolution is 1680x1050 I gotta put that into my original post
 
I dont have 3 monitors. Sorry.

My resolution is 1680x1050 I gotta put that into my original post

You need a bigger monitor man, several of them in fact, lol! 1680x1050 with a 7970 makes me :(

Thanks for the write up, I've been kicking around the idea of grabbing a couple 7950's over the 7970's.
 
You need a bigger monitor man, several of them in fact, lol! 1680x1050 with a 7970 makes me :(

Thanks for the write up, I've been kicking around the idea of grabbing a couple 7950's over the 7970's.

it may make you :( but it makes me :)

Its so hard to get these results. Because its multiplayer. So if I fly chopper in one, I wait to fly the chopper with the other card.

Im comparing these cards. For those who are reading this thread, I would insist that you look more at the minimum frames.

If anyone has a map or settings they would like me to try let me know.
 
Hey great benchmarks man, thanks for doing this. That sapphire 7950 OC is a beast of a card.

When the 7970 and 7950 are both at 1200/1700 there is like only a 3-4% difference
 
it may make you :( but it makes me :)

If anyone has a map or settings they would like me to try let me know.

Great comparo! Stuff like this really helps show what kind of $$ can be saved for most users out there.

Maybe try loading up on the AA a to strain them at bit more?
I would say most that are buying a 79xx would at least have one 1080p monitor, with many going the 120hz route where the 7970 might put them over the edge to get the sweet spot on a single display or the extra $100 isn't worth it for others.
 
Hey great benchmarks man, thanks for doing this. That sapphire 7950 OC is a beast of a card.

When the 7970 and 7950 are both at 1200/1700 there is like only a 3-4% difference

You are very welcome. Im about to test the 7970 in Gulf of Oman and Kharg Island.

With tessellation off there is pretty much NO difference between the two cards. With Tessellation ON, we will find out.

Question...where does BF3 use tessellation?
 
Great comparo! Stuff like this really helps show what kind of $$ can be saved for most users out there.

Maybe try loading up on the AA a to strain them at bit more?
I would say most that are buying a 79xx would at least have one 1080p monitor, with many going the 120hz route where the 7970 might put them over the edge to get the sweet spot on a single display or the extra $100 isn't worth it for others.

I plugged the 7970 back in and will go into Kharg Island and Gulf of Oman to test it with tessellation ON. I am using ultra settings for BF3.

I included AVP benchmark

Are there any benchmarks I can download to make this more fun? lol. I have a friends 6950 2gb unlocked into a 6970 2gb sitting here (have not given it back to him yet as I just received my card).
 
Everything is updated. Im tired. I hope this helps out.

Essentially, there is a slight difference between 7970 and 7950 in BF3 on a single monitor. If you get a card that clocks high, then youre in good hands. You cannot go wrong with either choice!

I hope this helps those who are on the fence about a 7970 vs 7950, personally I am sending my 7970 back. The coil whine is ridiculous.
 
This is why the HD7970 never interested me. HD7950 offers almost as much real world performance at a much better price.
 
Can you push the 7970 to the max? Or is the 1200 the max the reference cooler will handle?
 
Can you push the 7970 to the max? Or is the 1200 the max the reference cooler will handle?

Both cards are pretty close to their max as far as running BF3.

The 7950 could do 1250 on the core and 7970 1260-1270. BF3 is a big stressor on the cards, much larger than 3dmark, furmark, or anything else. A good session of bf3 will surely tell you if your card is unstable.

Both cards did 1200/1700 and I felt that its a great place to test. Some online review cards cant reach those speeds.
 
Damn. You're really making me regret my 7970 purchase.

Dont regret it. Is it a reference 7970? Does it clock high?

Keep in mind, my CPU at the moment is 4.2ghz and BF3 is the only game I tested. Ill gladly test more for you. I just dont have many games. Ill have the 7970 for another day before I ship it out.
 
I got the XFX dual fan black edition 7970. I figured that the BE is higher binned for more overclockability. It's in the mail at the moment, so I'll let you know how it clocks.
 
I got the XFX dual fan black edition 7970. I figured that the BE is higher binned for more overclockability. It's in the mail at the moment, so I'll let you know how it clocks.

You can decline the package and then get a full refund. Or you can keep the card. My conclusion is that you cannot go wrong with either card.
 
Do both 7 series at reference clocks so it gives a good baseline for your system. I love this thread though!
 
It's nice to see them both compared and appreciate the effort you went to do these benchmarks, but, I don't think you will get any idea of the difference between them at 1680*1050. Aren't you just running into CPU bottlenecks when the resolution is that low?
 
Really should be comparing these cards at a higher resolution, but great writeup anyways :)
 
Thanks for the hard work you put into this! Can't say I'm surprised though. The 7970 isn't going to pull away at this resolution.
 
Agreed. The resolution is too low to really test those 2 cards.
 
I don't think the resolution is too low. Both cards were maxed out at 99% gpu usage even at 1680x1050. I'll have an h100 next week so I can test it with increased CPU clocks. I would say for anyone wanting a card for 1920x1080 can't go wrong with either. I can't speak for eyefinity results.

It's tough to test online multiplayer but I tried to do similar things. Glad you guys like it. Unfortunately I don't have extra monitors.
 
the rez is too low to showcase the 7970's extra power. it won't start to pull ahead until you hit 30" monitor or eyefinity territory.
 
I don't think the resolution matters that much. Both cards have the same amount of VRAM so there's no issue there. Its basically just raw power at that point. A higher res might be useful but I doubt there would be more than a 10% difference between the cards.
 
I don't think the resolution matters that much. Both cards have the same amount of VRAM so there's no issue there. Its basically just raw power at that point. A higher res might be useful but I doubt there would be more than a 10% difference between the cards.

I agree
 
More results added. 7950 overclocked and regular speeds
 
Last edited:
Great review.

Just curious, do you have HBAO/SSAo on? Also, how much AA are you using?
 
2wiced, fantastic thread and writeup. I'm still on the fence between a 7970 and 7950, and these kind of comparisons are invaluable to me for this decision. Thanks for taking the time to do this!
 
2wiced, fantastic thread and writeup. I'm still on the fence between a 7970 and 7950, and these kind of comparisons are invaluable to me for this decision. Thanks for taking the time to do this!

You are very welcome. This forum has helped me. We all are a community and it just happened that I had an opportunity to help out. Im glad this provided you with helpful info.
 
nice write up. it makes the 7950 that much more appealing if the price comes down when kepler comes out
 
Picking up a 2600k and h100. I will redo some of the 7950 tests. Comparing 2600k. And then 4.2ghz vs 4.8ghz
 
Back
Top