BF 4: multi-player CPU bottlenecks

[beginner99]

Not sure if right forum or if I should use gaming forum. But here my question.

I don't own BF 4 yet but I've read about CPU bottlenecks even on high-end systems. These usually refer to 64-player maps filled with active players.

My question is: What about 32-player maps. Do these have the same issue? Or will does work OK even on older CPU's like lynnfield?

 

[OCGuy]

I think what you are hearing about was most likely the AMD marketing of their DirectX API competitor, Mantle. In order to show the greatest benefit over DX, AMD suggested it be tested in 64 player environments.

High-end systems already play the game so well, you won't notice a CPU bottleneck. Lynnfield plays the game just fine.

 

[MeldarthX]

depends really on your settings Beginner; video card; what you play at

 

[BrightCandle]

Admittedly my CPU is quite hefty but it runs 64 players BF4 at anywhere around 110-170 fps, and I am GPU limited the grand majority of the time (as shown by the in game tool) with 2x680's. Its not a particularly good example of a game that is CPU limited in all honesty.

 

[moonbogg]

Admittedly my CPU is quite hefty but it runs 64 players BF4 at anywhere around 110-170 fps, and I am GPU limited the grand majority of the time (as shown by the in game tool) with 2x680's. Its not a particularly good example of a game that is CPU limited in all honesty.

Agreed. BF4 just runs really well and uses the CPU really well. No bottlenecking that I have seen or noticed.

 

[beginner99]

Agreed. BF4 just runs really well and uses the CPU really well. No bottlenecking that I have seen or noticed.

But oyu guys have a 6-core and BF 4 is one of the few games that can make use of it?

depends really on your settings Beginner; video card; what you play at

I'm planning a display & GPU upgrade but if I'm bottlenecked by my CPU anyway I can save on GPU. I think I will go 1080p 120 hz. (at 1680x1050 now)

 

[moonbogg]

But oyu guys have a 6-core and BF 4 is one of the few games that can make use of it?



I'm planning a display & GPU upgrade but if I'm bottlenecked by my CPU anyway I can save on GPU. I think I will go 1080p 120 hz. (at 1680x1050 now)

What CPU do you have? What GPU were you thinking about getting?

 

[Grooveriding]

i7 is best for 64 player BF4, especially if you are chasing 120fps like you stated. You'll run into CPU bottlenecks on 64 player mp with an i5.

 

[beginner99]

What CPU do you have? What GPU were you thinking about getting?

i7-870 (@ 3.4 Ghz with turbo + HT enabled, could probably get higher OC. This is with stock voltage)

Unsure about GPU. probably R9 290(x) as better priced than NV (no mining craze here) and BF4 is included.

 

[OCGuy]

i7 is best for 64 player BF4, especially if you are chasing 120fps like you stated. You'll run into CPU bottlenecks on 64 player mp with an i5.

Link? The reviews everyone sites on this show HT providing no real benefit until you get down to i3.

 

[moonbogg]

Link? The reviews everyone sites on this show HT providing no real benefit until you get down to i3.

single or multiplayer? If single player then that chart is garbage.

 

[moonbogg]

i7-870 (@ 3.4 Ghz with turbo + HT enabled, could probably get higher OC. This is with stock voltage)

Unsure about GPU. probably R9 290(x) as better priced than NV (no mining craze here) and BF4 is included.

I think that combo would be great. You can use Mantle, so you won't be CPU bottlenecked anyway.

 

[OCGuy]

A lot of info given to people looking for help without any sort of data to back it up....very strange.

Case closed. HT simply doesn't provide any benefit in Battlefield 4 multi-player, regardless of whether you're using Win7 or Win8.1. We performed these tests on multiple maps and on multiple servers, and every time, the systems with HT enabled were slower. And yet, the use of Win8.1 definitely tightens up the measurements somewhat. The delta for average frames per second under Win7 was 9.5 percent, while it's only 3 percent under Win8.1.

Win 7

Win 8

http://www.techbuyersguru.com/ochtgaming4.php

 

[moonbogg]

A lot of info given to people looking for help without any sort of data to back it up....very strange.






Win 7

Win 8

http://www.techbuyersguru.com/ochtgaming4.php

Well I'm confused now. In any case, BF4 seems to use CPUs better than BF3 did so I think the OP will be fine with his i7 and a 290x, especially with Mantle helping his CPU out even more. I'd expect high FPS from that setup @ 1080p and a 120hz monitor will benefit from anything over 60fps, so he would be happy I'm sure.

 

[AtenRa]

Without Mantle, he will need to OC past 4.5GHz in order to be fine. I have my Core i7 3770K at 4441MHz with an HD7950 @ 1GHz. At 4GHz there were times that fps where going below 60fps. It all depends on the map and the number of players.
At Operation Locker he will be fine, in open maps like Siege Of Shanghai he will need to OC at least to 4.4-4.5GHz if he wants to stay at or above 60fps.

 

[Grooveriding]

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-china-rising-test-gpu.html



If you go i7 you'd see about a 30% increase over an i5 going after 120hz 1080p. i5 is just anemic for BF4 multiplayer. I see maxed out CPU usage at certain times in BF4 with a machine I have with a 4670K. Of course if you are running AMD, Mantle will eliminate the bottleneck most likely.

 

[Termie]

I'll keep this simple.

GameGPU has been publishing benchmarks since the beta showing that a CPU with Hyperthreading is 30% faster in BF4 than an equivalent CPU without it. No one else has found this, and despite my best efforts to replicate their findings, I've found practically the opposite to be true. Using HT is always slower, whether under Windows 7 or Windows 8, and whether using Ivy Bridge or Haswell.

 

[OCGuy]

Only on AT do people completely ignore data that resoundingly contradicts their opinion not two posts later. :biggrin:

 

[Grooveriding]

I'll keep this simple. GameGPU has been publishing benchmarks since the beta showing that a CPU with Hyperthreading is 30% faster in BF4 than an equivalent CPU without it. No one else has found this, and despite my best efforts to replicate their findings, I've found practically the opposite to be true. Using HT is always slower, whether under Windows 7 or Windows 8, and whether using Ivy Bridge or Haswell.

I would take a look at your system for existing issues. I don't get less performance with HT on.

The mutliplayer benchmarks are consistent; BF4 multiplayer gets significant gains from more cores and threads. Even the anemic AMD CPU is faster than the i5 4670K in the game. :ninja:


Gamegpu is one of the best review sites going. Them, computerbase.de and [H] put out the best game benchmarks going. Their results are consistent and reflect the reality of how these games perform on the hardware tested. Even their most recent benchmark of the Second Assault DLC adds up with less CPU gains seen in the more GPU-limited Metro 2014 map.

 

[moonbogg]

Conflicting observations and conflicting data. Hmm, I think i'll just keep my 3930k rocking with HT and forget about it.

 

[OCGuy]

Thats one way to read that chart...or you could see a 4c/4t i5 performing the same as 8c AMD @ 5ghz (wow!) and a 4c/8t i7. No idea why you would post that to back up the 4670k being "anemic".

 

[AtenRa]

Does anyone with a GTX 780Ti will play at 1080p and turn off Anisotropic Filtering and MSAA ????
Does anyone playing Multiplayer leaves Motion Blur ON ?? :whiste:
Even if you need 120fps, a 4GHz FX6300 will do the job at the same settings.

Those graphs are nice to get an idea of how the CPUs are performing. But the performance completely changes with different settings and different GPUs.
Same applies for every game.

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-second-assault-test-gpu.html

 

[CropDuster]

I turned ht off last night while playing and the game felt smoother to me. CPU usage peaked around 80% whereas with ht on its usually 35-40 iirc.

 

[Termie]

I turned ht off last night while playing and the game felt smoother to me. CPU usage peaked around 80% whereas with ht on its usually 35-40 iirc.

Thanks for testing this. The reason CPU usage doubled is that it's no longer counting the virtual cores, but as you discovered, those virtual cores do not work properly in BF4.

Others can continue to go by the GameGPU data, or they can test it themselves.

 

[Carfax83]

I never really noticed any difference myself with HT on or off myself. I think BF4 only uses the maximum eight threads in specific circumstances though, such as a fully loaded 64 player server..

Only then might HT be beneficial, as the CPU load would be very high. To me, the Frostbite 3 engine is well optimized for multicore CPUs..

When AMD and Dice published their Mantle benchmarks using the South China Sea mission, which according to Repi, is the most CPU intensive part of the SP campaign.

I replayed that same area, with my CPU at stock and overclocked settings, and there wasn't really much difference in performance which implies I'm not really CPU bound which I think is due to how effective the engine is at using my CPU cores.

I think I may play through that area again, this time with HT on and another with it off..