***THE BF4 BENCHMARKS THREAD***

[Twirlywoos]

Can't remember where I read someone who was having CPU spikes so I'll lob this in here:

http://answers.ea.com/t5/Battlefiel...emoryleak-PC-only-updated-May-13/td-p/1951455

Could it be a memory leak issue?

 

[LtMatt]

Can't remember where I read someone who was having CPU spikes so I'll lob this in here:

http://answers.ea.com/t5/Battlefiel...emoryleak-PC-only-updated-May-13/td-p/1951455

Could it be a memory leak issue?

I don't think that has anything to do with it, as i don't have a memory leak. This is some issue that has cropped up in the last couple of weeks for me.

 

[setter]

Downloaded that mate, on the main start screen where it asks for gpu details etc, what should be listed for file type and file name?

 

[LtMatt]

Downloaded that mate, on the main start screen where it asks for gpu details etc, what should be listed for file type and file name?

Like this mate.

 

[pgi947]

You can also use the BF4 frame time analyser (available via OCN) to give you a better idea of minimums. As a fraps read out may say that minimum frame was say 10fps, this tool will give you a percentage of how long was spent below a certain FPS to get a true idea of in game minimums like so:

It shows minimums of 30fps, yet spent enough time over that for it to be classed as rogue. Looking at the percentages, realistic minimums are more somewhere between 100-120fps as this is where the percentages start to show more frequent drops

Edit: Above reading was over a healthy 300,000+ frames from a lengthy 64 player conquest battle on flood CTE.

 

[LtMatt]

You can also use the BF4 frame time analyser (available via OCN) to give you a better idea of minimums. As a fraps read out may say that minimum frame was say 10fps, this tool will give you a percentage of how long was spent below a certain FPS to get a true idea of in game minimums like so:

It shows minimums of 30fps, yet spent enough time over that for it to be classed as rogue. Looking at the percentages, realistic minimums are more somewhere between 100-120fps as this is where the percentages start to show more frequent drops

Edit: Above reading was over a healthy 300,000+ frames from a lengthy 64 player conquest battle on flood CTE.

WHen i was getting cpu spikes my fps were dropping below 100 at times, where as prior to that i was up in the 175-200fps range. It was weird. I can't wait to check my results once this issue is fixed. Here's my results from that tool from oc.net - same results as i posted using FLA Calculator.



Don't really know what these results mean, but i do know they will improve a fair bit once those spikes are gone.

 

[pgi947]

The left hand column is your frame rate, the other 2 (I think) are the CPU and GPU I dividual frame rate, but not entirely sure what it means

But as you can see with your results, minimum of 52 yet spent 100% above that mark, your true minimum is somewhere between 120-144 as the percentage starts to drop a little over 144 - not by much though, mantle I presume? As one of it's selling points was more consistent frame rates with higher minimums which it shows.

Edit: its frame time calculates frame rates.

 

[LtMatt]

The left hand column is your frame rate, the other 2 (I think) are the CPU and GPU I dividual frame rate, but not entirely sure what it means

But as you can see with your results, minimum of 52 yet spent 100% above that mark, your true minimum is somewhere between 120-144 as the percentage starts to drop a little over 144 - not by much though, mantle I presume? As one of it's selling points was more consistent frame rates with higher minimums which it shows.

Edit: its frame time calculates frame rates.

Yeah its Mantle, but i was getting some serious performance drops. Its like every now and then there would be 5+ seconds of a spiky cpu line and fps would drop by 100 for a few seconds. Then it would shoot back up to 200fps. I'm pretty sure that has affected results so im curious to see what it produces when that doesn't happen. This never happened a few weeks ago, so something on Dices end is causing it.

 

[Twirlywoos]

GTX 670 SLI - 1280Mhz
i5 3570K - 4.5Ghz
Driver ver: 337.50

64 Player Conquest Zavod 311 - CTE

In Game Settings:

100 FOV

Spoiler

FLA Calculator

Spoiler

Frame Time Analyzer

Spoiler

User.cfg

Spoiler

WorldRender.MotionBlurEnable 0
WorldRender.MotionBlurQuality 0
WorldRender.MotionBlurMaxSampleCount 0
RenderDevice.RenderAheadLimit 3
RenderDevice.TripleBufferingEnable 1
RenderDevice.Dx11Enable 1
RenderDevice.Dx11Dot1Enable 1
RenderDevice.Dx11Dot1RuntimeEnable 1
Thread.MaxProcessorCount 4
WorldRender.TransparencyShadowmapsEnable 0
WorldRender.MotionBlurForceOn 0
WorldRender.MotionBlurFixedShutterTime 0
WorldRender.MotionBlurMax 0
WorldRender.SpotLightShadowmapEnable 0
WorldRender.SpotLightShadowmapResolution 256
WorldRender.LightTileCsPathEnable 0
WorldRender.PlanarReflectionEnable 0
PostProcess.DynamicAOEnable 0
GameTime.MaxVariableFPS 120

 

[setter]

Quick blast on zavod on normal bf4.

FLA C

Spoiler

Fta

 

[Twirlywoos]

Setter - Is that Ultra pre-set?

 

[setter]

It is mate, iirc its ultra, x2 msaa, (maybe 4x), 16xaf, post aa on high, HBAO on and motion blur off. 1920x1200. Will fire up bf4 just to check. Used to run sli 670's prior to this card in sig.

Was x2 msaa, post aa on high, turned the latter off and reset msaa to x4 now.

 

[LtMatt]

Question: What settings do people actually run with usually? Medium? High? Ultra? Customer with no AA???

1440P, Full Ultra +x4AA. I sometimes use resolution scaling as well as i have fps to play with.

 

[BaDBoY_uK]

So, does anyone feel that their settings have an influence - in any sort of way - on their ability to kill?

 

[Kaapstad]

So, does anyone feel that their settings have an influence - in any sort of way - on their ability to kill?

Makes no difference to me, I get slaughtered whatever settings I use.

 

[Rusty0611]

If I was going to get semi-serious about it I would turn settings down but BF has never really been that type of game. It's too noob friendly.

I wouldn't turn settings down for FPS but just to decrease screen noise (e.g. The foliage)

 

[whyscotty]

Makes no difference to me, I get slaughtered whatever settings I use.

Pretty much this

 

[LtMatt]

So, does anyone feel that their settings have an influence - in any sort of way - on their ability to kill?

Its easier to see enemies on low, but id rather play using Ultra settings to enjoy the eye candy as i don't play competitively or seriously.

Makes no difference to me, I get slaughtered whatever settings I use.

Pretty much this

Lol. Scotty has killed me a couple of times to be fair. Kaap won't even tell me his soldier name.

 

[humbug]

So, does anyone feel that their settings have an influence - in any sort of way - on their ability to kill?

In Crysis 3 MP on my old 7870XT if i turned the settings right down to keep high FPS i got a lot more kills for some reason, i don't know why, pointing my Gun at someone had them drop dead very noticeably more often than if i ran higher setting.

In BF3 and 4 that never made the blindest bit of difference, i run everything maxed in BF4, having said that i don't even remember what its like in DirectX which gives me much lower FPS, less than half in a lot of places.

 

[LtMatt]

In Crysis 3 MP on my old 7870XT if i turned the settings right down to keep high FPS i got a lot more kills for some reason, i don't know why, pointing my Gun at someone had them drop dead very noticeably more often than if i ran higher setting.

In BF3 and 4 that never made the blindest bit of difference, i run everything maxed in BF4, having said that i don't even remember what its like in DirectX which gives me much lower FPS, less than half in a lot of places.

Latency and i believe Mantle helps a bit with this as well.