• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

BF4 Retail CPU scaling measured

Ok some more results this time with my 3930k system (as per sig). Posted here and in the BF4 bench thread in case useful to people who don't check both sections of the forum.

Headline results with various CPU clocks below. I initially used my 24/7 GPU clocks and think it is fair to say there is little CPU bottlenecking going on from 4.5 GHz up to 4.9 GHz although min fps does improve as CPU clock goes up (if anything, there is GPU bottleneck). I then bumped the GPU clocks to 1200-1220 MHz just to see the effect and sure enough even at 4.5 GHz there is gain in average fps:






Below are some realtime data during the runs at the different CPU clocks. Quite a lot of thread utilisation but nowhere near all 12. GPU utilisation around 90% (for both). Click the images and then magnifying glass twice in bottom right to make bigger:

4.5 GHz


4.7 GHz


4.9 GHz


Will soon test multiplayer maps.
 
In SP
I am using 680 SLI at 1920x1080 ultra and there both running at around 90%.
CPU IVY 3770K 4.8Ghz is running at around 50%.
Clearly my VC's are maxed.
FPS bounces around between 90 - 120

If I use sync on refresh, (60hz) then GPU utilisation drops to around 70%.

Note: One card cant maintain 60FPS on it's own.
Two cards can and the SLI scaling is almost 100%


The heat coming out of my PC is unreal.
I swear I smelt burning plastic :)

Haven't run MP at all as I need to learn to play first :o
 
anyone seen any benchies comparing the 3570k overclocked to the 3770k overclocked?
how much improvement does the 4 virtual cores make?

all that i've seen so far compares them at stock (who runs k processors at stock?)
 
It seems Windows 8.1 improves performance a decent amount over Windows 7 in the game:

http://www.pcgameshardware.de/Battl...ts/Battlefield-4-Test-Benchmarks-CPU-1095298/

win 7 to win 8.0 yes little performance extra not much few fps does feel a little smoother . also there is no difference in performance from 8.0 - 8.1 this has been confirmed by repi on twitter head frostbite engineer.

so many people hyperboling fps benchmarks on i5 - i7 im so glad i didnt get a i5 why ? in all the proper benchmarks with singular card configs the fps difference in mp or single player is like 1 fps or same in some the i5 actually beats it in certain levels lol.

the benchmarks people keep showing often with i7s ;) hahaah are those with cards that are DUAL GPUS ! talk about bias benchmarks :rolleyes:

everyone knows in DUAL card set ups i7 is better ! we get that but making sweeping statements when in single card solutions which is what most game with or the majority its a bit of a bs move.

with top end cards likea titan or a 780gtx for eg no difference whatsoever with a i5 or a i7 ! 7990 is two cards in the benchmark people just been looking at lol.

best thing is ive tried the game and benched i5s against i7s so i know what fps is and what it isnt. atleast show the benchmarks fairly or dont bother.
 
So much BS and misinformation on public forums.

Max settings custom Ultra, 4xmsaa, 1920x1200p.


2x7950's (1200), 3750kx4.6Ghz, Max 85% Cpu ultilisation 1920x1200 VSYN ON...
60FPS Min, 60FPS Mean, Max 61 FPS..

You wont get a bottleneck even with 2x7950's with Vsyn on....(It might be occur at 120Mhz who knows!!)

I wont be upgrading my i5 3750k until I get my third 7950!!!!!


Battlefield 4 runs beautifully, but the damn game crashes ramdomly..


The game is heavy on the cpu but is bound by gpu's at higher refresh rates...
 
Last edited:
Without a built in bench it is hard to do comparisons.
I was thinking how to do a comparison between i5 and i7.

Using the same MB and swapping the CPU's is easy enough but finding a repeatable part of the game is more tricky.
 
So much BS and misinformation on public forums.

Max settings custom Ultra, 4xmsaa, 1920x1200p.


2x7950's (1200), 3750kx4.6Ghz, Max 85% Cpu ultilisation 1920x1200 VSYN ON...
60FPS Min, 60FPS Mean, Max 61 FPS..

You wont get a bottleneck even with 2x7950's with Vsyn on....(It might be occur at 120Mhz who knows!!)

I wont be upgrading my i5 3750k until I get my third 7950!!!!!


Battlefield 4 runs beautifully, but the damn game crashes ramdomly..


The game is heavy on the cpu but is bound by gpu's at higher refresh rates...

No offence, but there's a fair bit of misinformation in this post, and a few glaring technical inaccuracies.

It's Hz not Mhz (Given the massive difference)
On Vsync you're limiting the frame rate, so if there's a bottleneck you'd never know with Vsync turned on.

If you turned Vsync off, the bottleneck would be apparent whether or not it's a 120Hz/60Hz screen as you've unlocked it so it won't sync to the refresh rate.

Higher refresh rate limitation depends entirely on the set up (You could run a CPU that can only push 90 FPS, limiting your GPU set up capable of giving 130 FPS, or your CPU could be capable of 130, and your GPU's only 90 FPS)

We've seen in the Graphics Card forum that an i5 can bottleneck a Crossfire set up when you get into Multiplayer, an i5 isn't a foolproof solution for Multi-GPU.

Also, your resolution is higher than the average 1920x1080, it's not much higher, but it's going to lower your FPS compared to a 1920x1080 set up (Which the CPU would need to push)
What clock are your 7950's (People could be running higher clock 7970's capable of more FPS)
 
Last edited:
No offence, but there's a fair bit of misinformation in this post, and a few glaring technical inaccuracies.

It's Hz not Mhz (Given the massive difference)
On Vsync you're limiting the frame rate, so if there's a bottleneck you'd never know with Vsync turned on.

If you turned Vsync off, the bottleneck would be apparent whether or not it's a 120Hz/60Hz screen as you've unlocked it so it won't sync to the refresh rate.

Higher refresh rate limitation depends entirely on the set up.

We've seen in the Graphics Card forum that an i5 can bottleneck a Crossfire set up when you get into Multiplayer, an i5 isn't a foolproof solution for Multi-GPU.

+1

Turn vsync off and then keep an eye on the CPU/GPU usage.
 
breaking it down simple as possible

mp/single player one card set up no difference from i5-i7 regardless of settings.

mp/single player multi gpu set ups get a i7 its that simple.
 
No offence, but there's a fair bit of misinformation in this post, and a few glaring technical inaccuracies.

It's Hz not Mhz (Given the massive difference)
On Vsync you're limiting the frame rate, so if there's a bottleneck you'd never know with Vsync turned on.

If you turned Vsync off, the bottleneck would be apparent whether or not it's a 120Hz/60Hz screen as you've unlocked it so it won't sync to the refresh rate.

Higher refresh rate limitation depends entirely on the set up (You could run a CPU that can only push 90 FPS, limiting your GPU set up capable of giving 130 FPS, or your CPU could be capable of 130, and your GPU's only 90 FPS)

We've seen in the Graphics Card forum that an i5 can bottleneck a Crossfire set up when you get into Multiplayer, an i5 isn't a foolproof solution for Multi-GPU.

Also, your resolution is higher than the average 1920x1080, it's not much higher, but it's going to lower your FPS compared to a 1920x1080 set up (Which the CPU would need to push)
What clock are your 7950's (People could be running higher clock 7970's capable of more FPS)

USING MHZ INSTEAD HZ FOR MONITOR REFRESH RATE WAS CLEARLY A TYPO, YOU WERE A BIT ANAL TO REINTERATE IT ON YOUR POST.:)

Your stating the obvious above, that with vsyn off the gpu load goes up massively and hence too does the cpu usage. (Only around 5-7% in my case)

My point is that people can use a 60hz monitor at 60fps with 2x7950's with vsyn on and it shouldn't produce a bottleneck. This is different to stating that with vsyn off a bottleneck cannot occur, which I am not saying.

This topic is important because if people feel that the i5 wont work even at 60fps they will spend a lot of money upgrading to an i7 which for 60Hz gaming imo so far is not required.


There is no misinformation, I state once more that at 60hz/60fps with vsyn on there is not a cpu bottle neck with an i5 3750k and 2x7950's. (Running at 4.6Ghz and 1200Mhz respectively).

In my case with vsyn off I hit 170fps max, mins 80. Cpu 90% playing seige of shaghai 64 player

So with vsyn off things might be closer to becoming a bottleneck, but not at 60fps and 60hz.
 
Last edited:
You make some fair points, I agree with the sentiment, but I don't personally turn Vsync on (I like running Fifa at 400 FPS with single digit latencies)

The i5 can clearly perform perfectly fine in this game, I wasn't trying to argue that.
 
anyone seen any benchies comparing the 3570k overclocked to the 3770k overclocked?
how much improvement does the 4 virtual cores make?

all that i've seen so far compares them at stock (who runs k processors at stock?)

I cannot see any difference between a 4670k and 4770k both at 4.5ghz and a single GTX 780. Runs nice and smooth on either at 1080p or 1440p 60Hz.
 
You make some fair points, I agree with the sentiment, but I don't personally turn Vsync on (I like running Fifa at 400 FPS with single digit latencies)

The are a lot of games that are designed with V-SYNC in mind and will produce anomalies if the FPS goes over the optimal amount, that and OFC screen tear looks horrid.
 
The are a lot of games that are designed with V-SYNC in mind and will produce anomalies if the FPS goes over the optimal amount, that and OFC screen tear looks horrid.

I hate screen tear, but I don't ever get it in Fifa, even at 400FPS, I run a 120HZ monitor (Which should screen tear over 120HZ in theory)

In the past I've used Vsync, I remember Prince Of Persia wouldn't load if you didn't have Vsync on, so I get what you're saying about problems.
 
Back
Top Bottom