BF3 Multiplayer CPU scaling

[Caintz]

Hi,

there's a lot of discussing about this, was hoping for someone to have a benchmark (?):

How does the number of physical cores / threads scale for the FPS of Battlefield 3 in large multiplayer rounds? Benchmarks tend to rank Intel i3-2100 over AMD Bulldozers in single player mode, but what exactly happens in multiplayer- mode? Judging by forums, an AMD Phenom 1060T seems to outperform a dual-core Intel i3-2100 any time in a 64player map for example.

Second question: Is there a comparison for the FPS of i3-2100 vs. i5-2500k vs. i5-3550 in World of Warcraft?

Thanks in advance,
Caintz

 

[TheBigTroll]

CPU cores dont matter in games. the i3 2100 or 2120 will be more suited for your needs

cpu performance doesnt really affect games. all the cpu has to do in games is to prevent any bottlenecking on the graphics card. games can really only use 2 cores (except for bf3 where it can take up to 4 cores). the 1060t might outperform the i3 in some situations but it will use almost double the power doing so

i5 3550 (barely)>i5 2500k (if you overclock it, it will be better than i5 3550)>i3 2100

 

[Microgoliath]

Only in BF3 multi would a huge difference be seen when its a i3 vs phenom. I don't know exactly what happens and why but I would a assume

More players = more processing power needed.

On your second question, I couldn't really find a useful (on my phone) link but from personal knowledge WoW isn't very CPU dependent and won't differ much in performance when it comes to i3 vs i5 vs BD. That being said ofc the i5 will give you the most stable and best FPS in WoW / every other game.

 

[Caintz]

CPU cores dont matter in games. the i3 2100 or 2120 will be more suited for your needs

cpu performance doesnt really affect games. all the cpu has to do in games is to prevent any bottlenecking on the graphics card. games can really only use 2 cores (except for bf3 where it can take up to 4 cores). the 1060t might outperform the i3 in some situations but it will use almost double the power doing so

i5 3550 (barely)>i5 2500k (if you overclock it, it will be better than i5 3550)>i3 2100

I have thought so too for a very long time, but Phenom 1060T destroys i3-2100 in the 48-64 player multiplayer maps. . .

Thanks Microgoliath for a clarification

 

[kitsunestarwind]

BF3 in multiplayer will happily See improvement for up to 6 cpu cores for large 48-64 player maps, But it will not make use of Hyper threading at all

 

[gamerk316]

BF3 MP is VERY CPU intensive; the i3-2100 chokes to death in BF3 multiplayer.

 

[Caintz]

Is it impossible to request for a comparison from Tom's Hardware? I mean, a lot of people are wondering this but there is not a single benchmark of the multiplayer CPU scaling. . .

 

[amuffin]

Is it impossible to request for a comparison from Tom's Hardware? I mean, a lot of people are wondering this but there is not a single benchmark of the multiplayer CPU scaling. . .

BETA=no campaign, so always multiplayer

 

[Caintz]

BETA=no campaign, so always multiplayer
http://static.techspot.com/articles-info/448/bench/CPU_02.png

http://cdn.overclock.net/7/74/746236dc_2506.png

That still doesn't fit the question: how much does the amount of processor cores / threads affect performance in low / large multiplayer maps?

So basically am looking for a CPU comparation of a 16 player vs. 64 player map performance.

 

[-Fran-]

That still doesn't fit the question: how much does the amount of processor cores / threads affect performance in low / large multiplayer maps?

So basically am looking for a CPU comparation of a 16 player vs. 64 player map performance.

Well, it's really up to the Frostbite engine on how it processes all the physics and positioning/calculations in the MP maps.

I'm sure gamerk can correct me here, but no graphical engine is capable of handling what goes on in a multiplayer environment, so, whatever coding they did for the MP, it's quite different than SP.

In that regard, everyone with a 4+ core seems to have a better experience with MP maps in BF3.

Now, how much difference is there, it's one hell of a guesstimate, since there are a lot of considerations to make and I haven't seen any benchmarks comparing maps, just some MP. But you can trust this: the more people you see on your screen, the MORE your CPU will be taxed.

If you test an MP map with no people in it, then it's the same as testing a SP benchmark loop, so there's also that when looking the MP benchmarks.

Cheers!

EDIT: Only found this: http://www.hardocp.com/article/2011/11/22/battlefield_3_multiplayer_performance_iq_review/