Suprising Battlefield 3 CPU benchmarks

[swilli89]

http://www.techspot.com/review/448-battlefield-3-beta-performance/page7.html

I dont know if you guys have seen this preview but I was pretty surprised to see the CPU results. I guess it just means the game is near 100% which I expected but still.. Look at AMD's little A8 go! lol.

Interesting to me because I am wanting (needing actually) to replace my aging Conroe E6600 on my good ole badaxe 975x for BF3. I think my HD4890 may squeeze out a few more fps with a modern platform.

 

[TakeNoPrisoners]

Look at the x6 beating the i7.

 

[MiRai]

Look at the x6 beating the i7.

And the 920 is as fast as the 2600K...

I think we might need some non-beta benchmarks?

 

[TakeNoPrisoners]

And the 920 is as fast as the 2600K...

I think we might need some non-beta benchmarks?

Beating it by 2 fps means it still beat it and I ment the 2600k.

2fps is insignificant but the fact that it gets anywhere close to the intel CPU means the game must use all six threads, a good trend in my eyes.

The game is still not CPU bound so it doesnt mean much.

 

[swilli89]

Yeah it is a beta so could be a lot different once final game code is out.

What I don't want to see is reviewers show us "true" cpu results and benchmark at 1368x768 on Low settings to "remove the gpu bottleneck". I mean yeah it shows us something but its like a car mag taking off two wheels on a car and then racing them ."See which car drags its axle the best!?"

I want to see CPU's compared at normal resolutions with everything turned on.

 

[MiRai]

Beating it by 2 fps means it still beat it and I ment the 2600k.

2fps is insignificant but the fact that it gets anywhere close to the intel CPU means the game must use all six threads, a good trend in my eyes.

The game is still not CPU bound so it doesnt mean much.

I don't see how 3 extra FPS on the AMD system shows that 6 cores are being utilized over 4 cores but, I would guess that the chart shows the GTX 580 was able to put out 3FPS more in the AMD system as opposed to the Intel system.

It'd be great if you're right though, I'd love to see more games use my cores. From the conclusion of the review:

For a game that was seen utilizing six threads quite efficiently,

 

[Soccerman06]

Lol ultra settings rape everything, even sli 580s

 

[tulx]

Haha, Battlefield 3 is a monster. Might have to get another 5870 for this.

 

[WildW]

I was pondering an upgrade to Sandy Bridge, but it's starting to look like my Q8400 will cope just fine with BF3.

 

[cobra1981]

Looks like in the review it's a totally GPU limited situation and they're testing CPU scaling??? Doesn't make much sense to me. Why not take the resolution way down to 640x480, then test the CPU differences? This review doesn't show anything except that if you want to play on all high settings, get poor fps, then as long as you have a quad core you're good to go.

 

[bunnyfubbles]

Yeah it is a beta so could be a lot different once final game code is out.

What I don't want to see is reviewers show us "true" cpu results and benchmark at 1368x768 on Low settings to "remove the gpu bottleneck". I mean yeah it shows us something but its like a car mag taking off two wheels on a car and then racing them ."See which car drags its axle the best!?"

I want to see CPU's compared at normal resolutions with everything turned on.

that's fine if all you do is game with a GPU bottleneck

For some of us who like to use our processing power for other things while we game (such as recording video for later edit and upload to youtube or capture for live streaming) knowing which CPU is sweating the least is useful to know (as that processing power will be needed for live encoding), and the easiest way to do that is to remove the GPU bottleneck.

 

[-Slacker-]

Looks like in the review it's a totally GPU limited situation and they're testing CPU scaling??? Doesn't make much sense to me. Why not take the resolution way down to 640x480, then test the CPU differences? This review doesn't show anything except that if you want to play on all high settings, get poor fps, then as long as you have a quad core you're good to go.

Wouldn't it make sense that the performance difference between the tested cpus would only increase with lowering the resolution and visual details?

Also, why is it so important to see how hardware behaves in a scenario that no one will put the game through? Who plays BF3 at 640x480?

 

[3DVagabond]

Wouldn't it make sense that the performance difference between the tested cpus would only increase with lowering the resolution and visual details?

Also, why is it so important to see how hardware behaves in a scenario that no one will put the game through? Who plays BF3 at 640x480?

What he's saying is compare the 2 CPU's in a situation where the difference, if any, will be brought out in the benchmark. Not that it's real world, in any way. Most benchmarks aren't real world, really. If you are comparing the CPU's though lower res and let's see where the CPU pukes.

If the res is lowered and BD pulls 110fps and SB pulls 125fps though, who cares? What difference does it make? So, I see what you are saying.

 

[zlejedi]

Interesting I wonder what gives AMD edge here in a very unusuall way.

Also impressive result for i3 with HT compared to native dual cores.

 

[Vdubchaos]

Beta test means nothing.

They striped the crap out of this game, can't you tell?

It is weird that AMD "all of the sudden" has an edge in 1 game (out of 100s).

 

[-Slacker-]

What he's saying is compare the 2 CPU's in a situation where the difference, if any, will be brought out in the benchmark. Not that it's real world, in any way. Most benchmarks aren't real world, really. If you are comparing the CPU's though lower res and let's see where the CPU pukes.

If the res is lowered and BD pulls 110fps and SB pulls 125fps though, who cares? What difference does it make? So, I see what you are saying.

True, but the thing I don't understand is why would a processor have an edge at one resolution when it looses at another. Since cpu dependency would increase at lower resolutions, then wouldn't the amd processors lead only increase by a relatively large margin at 640x480?

 

[lol123]

True, but the thing I don't understand is why would a processor have an edge at one resolution when it looses at another. Since cpu dependency would increase at lower resolutions, then wouldn't the amd processors lead only increase by a relatively large margin at 640x480?

Yes. And that's the reason that CPU gaming tests are done at low resolutions (with the disclaimer that the large performance gaps in the results won't be reflected in real life scenarios), whereas the small differences you do get at higher resolutions are just due to noise.

 

[bunnyfubbles]

True, but the thing I don't understand is why would a processor have an edge at one resolution when it looses at another. Since cpu dependency would increase at lower resolutions, then wouldn't the amd processors lead only increase by a relatively large margin at 640x480?

because its likely that the so called edge in a GPU limited scenario is just margin of error and/or something else going on, whereas if we looked at CPU usage it could very well be that the X6 is being pushed to 80% while the 2600K might be only around 60%

for a negligible amount of frame difference the better CPU would clearly be the one with lower usage and have far more headroom

 

[Genx87]

Looking at how little difference there is between the processors. It looks like the benchmark is GPU limited.

 

[Ferzerp]

It's time for all these reviewers to start applying at least rudimentary statistical methods if they're going to start trying to review processors in GPU bound situations. We need 100 or so runs per set up with the mean and the standard deviation before values this close mean anything at all.

 

[Maximilian]

Looking at how little difference there is between the processors. It looks like the benchmark is GPU limited.

This.

 

[xsilver]

Anybody have a link to how a regular core 2 duo performs on a more modest card? eg. 6850/6870? Would it be cpu or gpu limited?

 

[bunnyfubbles]

it doesn't take rocket surgery to figure out BF3 will be first and foremost GPU limited, although the more interesting CPU tests will come in the heavily scripted single player scenarios as well as the large 64 player maps with tons of vehicles and jets flying all over the place

 

[nsdjoe]

Anybody have a link to how a regular core 2 duo performs on a more modest card? eg. 6850/6870? Would it be cpu or gpu limited?

I have a 6870 but I can tell you my Core 2 Quad (Q6600) was definitely bottlenecking my performance, at least while playing 64 player Caspian Border, where I often dipped to 20-25fps (unplayable imo). 32 player Operation Metro was completely fine, however.

 

[xsilver]

I have a 6870 but I can tell you my Core 2 Quad (Q6600) was definitely bottlenecking my performance, at least while playing 64 player Caspian Border, where I often dipped to 20-25fps (unplayable imo). 32 player Operation Metro was completely fine, however.

@1080p?
ultra settings?
dont mind playing on high and only running 1440x900