I just thought I'd post this because it may be useful to some people who are wondering about this game's "Crysis 1" style beefiness. Also, I haven't seen many results floating around the web, and probably because the game is still hella buggy and new. Regardless, this will be much more realistic compared to the Alpha and Beta benchmark results on the interweb. I hope this helps people out!
Here's a quick rundown of what I did. I ran 4 tests with Fraps in BF4 in a controlled environment for two minutes each run. I know that's where the advantage is, but it's the best way to get a "best possible frame rate," without a whole lot going on. I ran these tests on the level where Recker wakes up and hops out of his bunk aboard the USS Valkyrie. The ship has a lot of HDR lighting, reflective textures, water, people, shadows, and even a carrier on fire! To sum it all up, it has all the basic visual effects that you can expect in the game, with exception of shooting and full on explosions.
Before I waste anyone's time who thinks this is a FULL review/benchmark from every setting starting from low to ultra, it's not. It's strictly a spectrum difference between low and ultra. It would be extremely time consuming to do everything, and the game's performance will probably change once patches and drivers start rolling out. Please take this for what it's worth, especially if you're in a position where you have to sacrifice the eye candy. For the record, high and medium net about the same results with +-5FPS difference.
The system that I used isn't uber-high end. I would be honest and say it is a mid-to-high-end system; it is something that an enthusiast would use for PC gaming. Unfortunately, I wasn't able to monitor the temperatures before, during, and after the tests. Here are the specs:
AMD FX-8350 stock clock/cooler
ASRock Extreme4 970 (BIOS P2.30)
Corsair Vengeance 16GB DDR3-1866
HIS HD 7950 IceQ Boost (AMD Catalyst 13.11 1100Mhz Core/1380Mhz Mem/+10%Power)
Hiper 880W Type-M PSU
Samsung SP200 200GB HDD
3x Samsung Syncmaster 906BW (4320x900/4536x900 Bezel Comp)
For the record, I did this test with Eyefinity. It would have made more sense if I did a single monitor benchmark. Just FYI though, I did the same tests on a single display, but forgot to make any record of it because it was on the South China level where you have to destroy the tank. I just want to make that clear. I also want to add that single screen performance was surprisingly great.
The first run was at 4536x900 on Low with MSAA off and HBAO/SSAO off
Avg: 33.133 - Min: 21 - Max: 46
I actually think that this is pretty playable. There isn't much lag at all between mouse movement and response, but it is still noticeable yet negligible. I find myself playing this setting simply because it's the smoothest out of all, and yet the game still looks pretty decent. Multiplayer yields around the same results.
Second run was at the same resolution on Low with 4x MSAA and HBAO on
Avg: 27.650 - Min: 20 - Max: 36
This is actually not too bad either. I don't know if the OC speeds on the GPU are helping out, or the 3GB frame buffer, but it is still extremely smooth for having these settings. I wouldn't play with them on because the textures look dull but the edges look crisp. I wouldn't mind playing the other way around. For bench sakes and time, we'll just go with textures on low since that kills FPS at this resolution, believe it or not.
Third run, same resolution on Ultra with MSAA off and HBAO/SSAO off
Avg: 11.683 - Min: 9 - Max: 15
This is barely playable to say at the least. It's borderline slide show, but it just looks soo good! I don't recommend playing this setting at all, unless you want to enjoy a slow first person shooter. The mouse response is definitely effected by this. This sample was run between the part where the tug was towing the V-22 onto the elevator through the part where the USS Titan was blown up. This sample has the water, HDR, reflections, smoke, fire, textures, etc.
Last run, same resolution on max settings (Ultra, 4xMSAA, HBAO)
Avg: 10.658 - Min: 6 - Max: 17
This is unacceptable for me or anyone else. I don't know what else to say about this. The game looked great. I had this setting on the game when you meet Garrison and a little shortly before. There is impeccable detail in the weaponry and accessories that your team is wearing. If I had a powerful system, I would run it at this setting every time.
Now, I know this thread is worthless without pics...so for all of those people who read benchmarks and reviews for the charts and numbers, I decided to format something for all of you (and for me) to see the differences and see where things are with colors, and numbers, and lines, oh my!
I hope this will help anyone make a more informed decision on an upgrade. I'm now more convinced about getting another HD 7950 and a good cooler for the CPU. I'm sure a well balanced system would have no problem with this. I'm hoping that the CrossFire HD 7950 will at least double these results, coupled with a strong and stable CPU overclock. If you read this whole thing and made up to this point, I apologize for posting a some what arbitrary benchmark. Hopefully everyone here who read this is able to make a more informed decision on surround gaming or your next upgrade. All in all, I'm actually extremely satisfied with the results I got, but I'm hoping that the CrossFire and higher clocks yield much better results! In case you were trying to figure out whether or not you should upgrade your GPU or just throw another one in there depending on which one you have, here's this nifty tool that I found where I compared two 7950's against a GTX690. I'm certain that a combination of other cards can show greater improvements, whether you have two GTX 660 Ti, GTX 760, 7870, 7890, or a single GTX 780. There will always be issues with multi-GPU solutions, but it would probably be worth the headache over the one you receive after spending $1000 on a single GPU