BF3 CPU bottleneck? Really?

[moonbogg]

OK,
I have a 2600k @ 4.3 with HT on. Two GTX 670's in SLI and 8 gigs system ram @ 1866. I play BF3 a lot and care about high FPS because I use a 120hz 1080p monitor.
I see what seems like wicked CPU limitation, but I am not experienced enough to really diagnose this without help. On almost all maps I notice wildly fluctuating GPU loading from 50%-95%. It is usually around 60-70%. I also watched CPU usage and it was (according to the software) about 60-70% across all 8 threads.
My FPS is usually between 80-120, but on B2K maps it goes to 55 or so at some points. This is a widespread "problem" that a lot of people experience with dual high end card setups.
I know the 2600K is fast, but that doesn't mean it isn't bottlenecking. How can I test and measure to actually know what is happening? Should I try to OC my CPU more? It already hits 60c under max load. I might consider a 6 core CPU, but that might be ridiculous and not yield any significant benefit. I heard BF3 fully uses 8 threads so I duno, maybe it would help.

 

[Termie]

I can pretty much assure you that minimums in BF3 with that much GPU power are being dictated by your CPU. Furthermore, at that resolution, even your averages are going to be severely limited by your CPU.

Anandtech found BF3 scaling to be around 64% on its 680SLI setup at 1920x1200, FXAA: http://www.anandtech.com/bench/Product/585?vs=555.

This was in single-player. In multiplayer, it would likely be worse. Basically, you have too much GPU power for that game, which is heavily dependent on CPU, especially for the B2K maps, with their enhanced destruction. You just aren't going to be hitting 120fps constant as you probably hoped.

 

[blckgrffn]

Time to go LGA 2011, man! :awe:

 

[moonbogg]

Yeah that makes sense and shouldn't be suprising at 1080p. Don't get me wrong though, SLI provides much higher FPS than a single card, its just that I can see some CPU limitations. What bothers me even more is that if a 2600k @ 4.3 can't handle the game at 60fps then what the hell can?

 

[T_Yamamoto]

4.3ghz? TOOSLOW

 

[Rvenger]

Yeah that makes sense and shouldn't be suprising at 1080p. Don't get me wrong though, SLI provides much higher FPS than a single card, its just that I can see some CPU limitations. What bothers me even more is that if a 2600k @ 4.3 can't handle the game at 60fps then what the hell can?

Your minimums shouldn't hit below 60 for a 2600k and around 55 with HT off. This has been proved over at OCN. (I forget the thread) Something else is not working right in your rig.

 

[moonbogg]

Your minimums shouldn't hit below 60 for a 2600k and around 55 with HT off. This has been proved over at OCN. (I forget the thread) Something else is not working right in your rig.

I think the rig is working just fine. I am simply expecting more out of it than what it can do. my only option might be to upgrade platforms, but its just not worth it for an extra 15fps of min framerate or so. Most of the time its really high so screw it. That 2600K is showing its age. We need more IPC damn it.

 

[felang]

I think the rig is working just fine. I am simply expecting more out of it than what it can do. my only option might be to upgrade platforms, but its just not worth it for an extra 15fps of min framerate or so. Most of the time its really high so screw it. That 2600K is showing its age. We need more IPC damn it.

Just for the hell of it, have you tried to disable HT just to see how high the CPU usage goes? I´m planning on getting a second GTX 680 and also have a 2600K, but want to see if it´s worth it or not. Although I would be using 1440p rez but only 60hz.

Also, you should be able to easily get at least 4.5 Ghz from a 2600K at 1.3V or less. If you have decent cooling go to at least 1.35V (maybe even 1.38-1.4) and see how high you can clock that b*tch.

 

[pelov]

I think the rig is working just fine. I am simply expecting more out of it than what it can do. my only option might be to upgrade platforms, but its just not worth it for an extra 15fps of min framerate or so. Most of the time its really high so screw it. That 2600K is showing its age. We need more IPC damn it.

You're not going to gain anything from 2011 that you don't already have on 1155 if we're talking gaming performance. As far as gaming goes they perform the same.

Furthermore, have you checked your CPU utilization during gaming? This definitely does not sound like a CPU bottleneck but more likely something else being the culprit.

 

[Kevmanw430]

Dare I say it, 4.3GHz is almost conservative. You could probobly get another 500 MHz out of it if you have decent cooling.

 

[Don Karnage]

A 4.3Ghz 2600K a bottleneck? Not even close. Up your overclock to 4.6Ghz+

 

[Black96ws6]

I feel like I'm just repeating what everyone else has already said, but add my 2 cents to this:

1. 4.3Ghz is low for a Sandy Bridge chip. You need to bump that up to 4.5Ghz+ if you are on anything other than the stock Intel cooler!

2. When you look at CPU utilization in Windows, you will not see CPU usage @ 100% (unless your CPU is getting hammered) because it splits the load evenly across all the cores.

3. Do you have the latest drivers from Nvidia? They're always releasing updates that improve SLI performance, so make sure you have the absolute latest.

I'm sure there's more I'm forgetting however start with increasing your overclock and ensuring your drivers are up to date. Please note overclock at your own risk (lol my disclaimer .

 

[T_Yamamoto]

A 4.3Ghz 2600K a bottleneck? Not even close. Up your overclock to 4.6Ghz+

like i said. TOOSLOWOC

 

[VirtualLarry]

Just for the hell of it, have you tried to disable HT just to see how high the CPU usage goes?

Do this. With HT enabled, you won't see a true accounting of the CPU usage.

 

[Prey2big]

I would advice you to keep hyper-threading on. It will surely give a healthy boost in BF3 multiplayer, especially during intense combat on 64-player servers.
It most definately is in BF Bad Company 2 anyways.

But you can turn it off momentarily for measuring purposes.
Let us know what the CPU graphs look like

 

[Don Karnage]

I would advice you to keep hyper-threading on. It will surely give a healthy boost in BF3 multiplayer, especially during intense combat on 64-player servers.
It most definately is in BF Bad Company 2 anyways.

But you can turn it off momentarily for measuring purposes.
Let us know what the CPU graphs look like

Games dont utilize Hyperthreading

 

[Termie]

Games dont utilize Hyperthreading

Yes they do. And BF3 can take advantage of it.

 

[BD231]

GPU's are way faster than CPU's (paralleled work loads), so yeah if you hook two into one CPU obviously those 2 gpus will be pushing out more frames than your cpu can handle in a lot of situations.

Your CPU is not the issue though. BF3 is one of the worst, if not *THE* worst game you'll ever play when it comes to scaling in cross fire or SLi, it just plain sucks. Perhaps CPU physics play a part but you see the same issue on low end SLI/CrossFire setups as well so it's definitely not entirely your CPU. It's the game.

 

[Don Karnage]

Yes they do. And BF3 can take advantage of it.

Added cache and the 100mhz clock speed advantage is why it's faster in BF3

 

[pelov]

That graph really reaffirms the point that it's not the CPU that's causing the problems here. When the FX4100 performs as well as a 2600K it shows just how GPU heavy that game is even at 1680x1050 resolution.

Looks like it's this

Your CPU is not the issue though. BF3 is one of the worst, if not *THE* worst game you'll ever play when it comes to scaling in cross fire or SLi, it just plain sucks. Perhaps CPU physics play a part but you see the same issue on low end SLI/CrossFire setups as well so it's definitely not entirely your CPU. It's the game.

 

[Termie]

Gents, if you've played BF3 multiplayer, you should know how this works. It's heavily multi-threaded and it's heavily dependent on CPU. Throwing up singleplayer benchmarks from BF3 is about as relevant as showing me a graph from Metro 2033.

Here's the proof:

Back to the OP's question, while disabling HT may help you figure out the CPU load, it surely won't help increase your minimums. Can you give us a FRAPS run from Oman so we can get an idea for min and averages to tell if there's actually something wrong?

 

[Don Karnage]

Gents, if you've played BF3 multiplayer, you should know how this works. It's heavily multi-threaded and it's heavily dependent on CPU. Throwing up singleplayer benchmarks from BF3 is about as relevant as showing me a graph from Metro 2033.

Here's the proof:

Back to the OP's question, while disabling HT may help you figure out the CPU load, it surely won't help increase your minimums. Can you give us a FRAPS run from Oman so we can get an idea for min and averages to tell if there's actually something wrong?

Whats with the joke graph? Medium settings?

 

[skipsneeky2]

I honestly believe dual gtx670 cards are bored at 1080p in BF3 hence the different gpu usage readings as i got dual stock clocked gtx670s but a 3770k@ 4.2ghz and my frames are anywhere from 67-180 with complete maxed settings with nearly the same readings as the Op.

Running a 60hz t.v but soon to be a 120hz so not sure if that matters.

 

[SickBeast]

4.8ghz or ban! *With* HT enabled!!

Moar power! Moar clock speed! Moar frames per second! Moar moar moar!!

 

[Carlitos714]

You need to overclock you cpu some more.
1920 X 1200
i7-920 sli gtx 480's
I only see 100% usage on my gpu's when Im overclocked to at least 4.1ghz.
You have better gpu's so you need to overclock that cpu more. Post results when u do!