***THE BF4 BENCHMARKS THREAD***

[Rusty0611]

Aside from the links already provided and a ton of feedback from people using 2gb cards on various forums around the web and on battlelog. (yes i know )

I have here some benchmarks where the 2gb limit was not exceeded but resulted in the problems i mentioned above. Given the demands of bf4 its easily possible to see this behaviour happening again with much higher vram usage.

HardOCP also posted something exactly like this regarding Hitman Absolution. Can't remember the link but it showed exactly the same as Swe clockers and pcper's conclusion.

Yes, but look at the FPS though - it's nearly identical...

As for the frame time comparison - after I was told my 7950 CF set up was unplayable before due to frame times and it was fine I tend not to really give it much time .

Again, they offer no technical reasons as to why the difference is there. It's guess work.

 

[Silent_Scone]

And if you don't think a 2GB 680 is hitting limits at 4K you're barking mad!

I don't comprehend... can you elaborate?

You mention issues with 8XMSAA, on what card? That'll turn a few cards into a slide show. Combination of lack of grunt and yes, VRAM.

 

[Rusty0611]

And if you don't think a 2GB 680 is hitting limits at 4K you're barking mad!

Why's the FPS the same then as the 4GB version?

 

[LtMatt]

Yes, but look at the FPS though - it's nearly identical...

As for the frame time comparison - after I was told my 7950 CF set up was unplayable before due to frame times and it was fine I tend not to really give it much time .

Again, they offer no technical reasons as to why the difference is there. It's guess work.

But my point all along has been, it can manifest itself in different ways. Fps drops, stuttering, hitching, decreased performance. It may also be as you say if you greatly exceed the limit bang 1fps. Its clearly not always going to be the same way. The numbers speak for themselves. Swe clockers have no reason to lie and make up those numbers.

 

[pgi947]

Going round in cicles here guys. Agree to disagree?

If folk actually bothered to send me all their results perhaps we could get all the results side by side and actually see whats what?

 

[Rusty0611]

But my point all along has been, it can manifest itself in different ways. Fps drops, stuttering, hitching, decreased performance. It may also be as you say if you greatly exceed the limit bang 1fps. Its clearly not always going to be the same way. The numbers speak for themselves.

The numbers don't say the same thing you're saying though.

You haven't answered why a single 770 is achieving a higher minimum than a 690. The answer is very simple: game is unoptimised and it's really difficult to draw out general conclusions like "2GB is not enough".

Kaap has already said he's maxed out at 1600p with 2 690s and got a minimum of 60 odd....

 

[flopper]

cpu matters, I would like to see,
4ghz 4 core without HT
4ghz 6 core without HT
4ghz 4core with HT
4ghz 6 core with HT
Or such on 64 map Multiplayer.
Both amd and Intel.
Mantle might make a huge difference down the line for amd.

Beta runs ok, better than Bf3 did for me
more fun gameplay makes a huge difference.

 

[Silent_Scone]

Why's the FPS the same then as the 4GB version?

Refer to HARDOCPs findings.

 

[LtMatt]

Going round in cicles here guys. Agree to disagree?

If folk actually bothered to send me all their results perhaps we could get all the results side by side and actually see whats what?

One thing to consider. If you run 32 man domination you get much better fps and lower vram usage. If you run 64 man conquest large you get lower fps and higher vram usage. You should probably ask which mode/size people have played on as according to my testing it can make a big difference to the results.

 

[sand_dune]

Why's the FPS the same then as the 4GB version?

Just read the graph dude. It's the same for you and me. You can see that fps is about the same, while one card clearly runs out of memory at times. And back up mechanism kicks in and you need more Vram!

 

[Mei]

But my point all along has been, it can manifest itself in different ways. Fps drops, stuttering, hitching, decreased performance. It may also be as you say if you greatly exceed the limit bang 1fps. Its clearly not always going to be the same way. The numbers speak for themselves. Swe clockers have no reason to lie and make up those numbers.

ye i dont buy that *you would see 5fps flickbook style* if run out of vram
i think it can not run at its most optimum without coming to a grinding halt

 

[Rusty0611]

Refer to HARDOCPs findings.

They don't offer any technical reasons as to why the 2GB produces marginally worse frame times other than a guess at it being a 2GB card. It doesn't even constitute a theory.

The frame rate itself is nigh on identical.

 

[Jono8]

Uncle Petey with his 7970.

Frames, Time (ms), Min, Max, Avg
51837, 737463, 7, 129, 70.291

Vram usage: 1998Mb

All Ultra with 2xMSAA

2500K @4.5
7970 @1230/1800

The min fps is down to some kind of random glitch with the beta in my opinion.
Once or twice during a round, it'll stutter for a couple of seconds whether I'm surrounded by explosions or just walking down the street with absolutely nothing going on.

Generally, the fraps counter (if I'm not benching) will show 55-80 fps, regardless of the situation.

Maybe you are running out of vram

 

[Silent_Scone]

You can't just go to afterburner and go, ah look that's only using 1840. All is well.

Mean while your drivers are shovelling in the furnace trying to keep on top of all the wonderous things you see before you on your screen.

It's not all about the frame times!

 

[Gregster]

I will fire over an email now Paul. Sorry for my part in ruining this thread.

 

[LtMatt]

The numbers don't say the same thing you're saying though.

You haven't answered why a single 770 is achieving a higher minimum than a 690. The answer is very simple: game is unoptimised and it's really difficult to draw out general conclusions like "2GB is not enough".

Kaap has already said he's maxed out at 1600p with 2 690s and got a minimum of 60 odd....

Its hardly surprising. If both cards are tanking because of the frame buffer which would provide a 1-2 higher minimum fps? Single card or dual card? Single card most likely. The 770'd faster memory chips could account for this difference.

Kaap said and kaap showing it are two different things. You get lower vram usage and more fps depending on game mode played and map size used. I've tested this myself. There was no mention of what settings Kaap used. Last time i saw Kaap with 70FOV at 1600p for christs sake. Kaap was also saying not long ago his 690's were on 20 fps and SLI was not working. Now its working and he's got 60 minimum for four way SLI. No offence Kaap.

 

[LtMatt]

Maybe you are running out of vram

Did you miss the part at the top Jono? Those were Petey's benchmarks playing on a laggy server.

 

[Mei]

don't take the lords name in vain!

 

[Rusty0611]

You can't just go to afterburner and go, ah look that's only using 1840. All is well.

Mean while your drivers are shovelling in the furnace trying to keep on top of all the wonderous things you see before you on your screen.

It's not all about the frame times!

No offence but I don't think you know what you're talking about. As I said, unless you care to elaborate with a technical explantion?

Its hardly surprising. If both cards are tanking because of the frame buffer which would provide a 1-2 higher minimum fps? Single card or dual card? Single card most likely. The 770'd faster memory chips could account for this difference.

Kaap said and kaap showing it are two different things. You get lower vram usage and more fps depending on game mode played and map size used. I've tested this myself. There was no mention of what settings Kaap used. Last time i saw Kaap with 70FOV at 1600p for christs sake. Kaap was also saying not long ago his 690's were on 20 fps and SLI was not working. Now its working and he's got 60 minimum for four way SLI. No offence Kaap.

Doesn't sound likely. It's not that complex. The game just isn't optimised for us nVidia bums . Kaap doesn't need to show it - I don't think he's lying. He already said that SLI wasn't working because he was in windowed mode. Switching to full screen fixed it.

 

[Jono8]

I'm siding with the tech site who have done proper benchmarks on all the cards at all the settings. The stark difference between the 690/770SLI results here tells you the 2gb buffer is struggling. You can see the low minimums on Ultra with high vram usage, vs medium settings with low vram usage while the minimum fps are much better as the buffer is not being filled.

Swe clockers come to the same conclusion i do. And so do all of their user comments.

How do you explain the 7950 boost and 760 results then? The 760 has higher minimum and average framerates : /