7970 X 680 @ 2560 X 1600 BF3

Locky · 25 Mar 2012 at 00:59

Peaked @ 1660 for me. Kiwi was a bit lower.

Surveyor · 25 Mar 2012 at 02:12

rare said:
Bizare how the 680 is peaking when the 7970 is dropping off?

I wish I'd spotted that.

Oh wait, I did

rare · 25 Mar 2012 at 02:32

I know, sorry..... Just trying to reemphasise the point. Very confusing??

1kiwi · 25 Mar 2012 at 13:00

Indeed they are out of sync at 55 and 91. We were trying to keep in sync as much as possible but there were times when I died (I do that allot in BF3) and Locky kept playing.

Just thought it would be interesting to try and do some real world tests unstead of operation swordbreaker

Cythx · 25 Mar 2012 at 13:06

kiwi1 said:
Indeed they are out of sync at 55 and 91. We were trying to keep in sync as much as possible but there were times when I died (I do that allot in BF3) and Locky kept playing.

Just thought it would be interesting to try and do some real world tests unstead of operation swordbreaker

4m of 'real world' leaves a lot of room for error.

Helios1234 · 25 Mar 2012 at 13:27

PanMaster said:
Battlefield 3 doesn't make for a very useful comparison test on its own being a console game and disliked by many including me. Dirt 3, Crysis2 and Skyrim (no mods) are just as useless.

Are there any non-biased niche games such as Egosoft's X-Rebirth, the Anno series, various RTS, Skyrim with heavy mods, that someone could run a comparison on? edit: Crysis (not Crysis 2 which is crap) would be nice too.

So that makes the most popular games currently out on PC useless for benchmarking? notsureifserious.jpg

555BUK · 25 Mar 2012 at 13:53

Saying that the best selling PC game of it's genre is not a very useful comparison test is probably a little shortsighted. As >75% of current PC games are now console ports, we really wouldn't be left with many valid game tests.

The best selling games of 2011/2012 are clearly the most valid games to test. BF3 is probably the foremost of these.

Gregster · 25 Mar 2012 at 13:57

I personally found BF3 a good game to bench as it pushes the GPU fully. Thanks for showing us the comparison Kiwi and Locky.

beejjacobs · 25 Mar 2012 at 14:00

gregster said:
I personally found BF3 a good game to bench as it pushes the GPU fully. Thanks for showing us the comparison Kiwi and Locky.

I agree, and plenty of people play it/will buy GPUs for it.

Cheers for putting this together, nice to have some 'real world' comparison!

PGdude · 25 Mar 2012 at 14:06

Thanks for the comparison guys.

Do either of you have Shogun 2? Might be easier to do a consistent bench of there

1kiwi · 25 Mar 2012 at 14:41

No worries guys. We are working on a better test for a longer duration. I just need to work out what is the best stable clock for me. Prob around +160, +450.

Locky · 25 Mar 2012 at 14:58

I don't think my 7970 was working correctly so, we are going to rerun the test, ignore the one in the op, an updated one will but up shortly.

Getting average of 55fps now and 3dmark 11 is 10.4k

1kiwi · 25 Mar 2012 at 18:26

OP updated.

Surveyor · 25 Mar 2012 at 18:37

kiwi1 said:
OP updated.

Changing the colours around on the two graphs doesn't help.

Overall it looks like the HD 7970 gives the best performance.

Cupra · 25 Mar 2012 at 18:38

Surveyor said:
It still looks very odd.

The HD 7970 seems to give better frame rates until you die then it takes a nose dive.

the 680 seems to be a little better most of the time, are you sure ur reading it correctly?

Locky · 25 Mar 2012 at 18:39

Surveyor said:
Changing the colours around on the two graphs doesn't help.

Yeah It doesn't help, cba rebuilding the graph

Surveyor · 25 Mar 2012 at 18:42

Cupra said:
the 680 seems to be a little better most of the time, are you sure ur reading it correctly?

In the second graph red is now the HD 7970.

It seems to be above the blue line of the GTX 680 more often than not.

It's not that easy to interpret.

Surveyor · 25 Mar 2012 at 19:20

Locky said:
Yeah It doesn't help, cba rebuilding the graph

It seems to have confused Cupra

555BUK · 25 Mar 2012 at 19:38

The second graph looks pretty equal to me. The 680 looks better for the first half of the test, and then in the second half the 7970 looks better. It's hard to choose between them.

Locky · 25 Mar 2012 at 19:48

555BUK said:
The second graph looks pretty equal to me. The 680 looks better for the first half of the test, and then in the second half the 7970 looks better. It's hard to choose between them.

There was like 1fps better average in favour of the 7970 in reality a you can't get a fag paper between them.