Win8.1 BF4 performance

[omeds]

I saw the troll thread was rightfully locked, however its an interesting topic and this isn't about mantle so I think it deserves it's own thread.

As some of you may or may not know, Nvidia seems to be leveraging DX features under win8.1 which offer some similar benefits as Mantle in BF4 - reduced draw call and CPU overhead. For whatever reason, AMD is not doing the same (probably because they figure everyone will use Mantle, and they've always had trouble with DX multithreading)

Discuss

 

[itsmydamnation]

see the word beta..... yeah... this thread needs to closed..........

 

[biostud]

Old news:
http://www.hardocp.com/article/2013...ows_7_vs_81_performance_review/4#.Uu99s7SwRrk


results when not in Beta:

 

[omeds]

see the word beta..... yeah... this thread needs to closed..........

What about it? The beta also gained performance increases in 8.1.

The results in the second graph are from only 2 days ago.

 

[Gloomy]

This has been the case for a while, lol.

This last patch also increased Nvidia's performance by a lot.

 

[omeds]

Yes I noticed this too, the recent [H] GHz review in 8.1 shows it dominating the 290 under DX. Then, within the same week Pclab runs the GHz under 8.1 and get very similar results:

Are there any other games that exhibit this behavior?

 

[biostud]

I would rather have Mantle review with 780 and 780Ti scores included from one of the major sites.

 

[blackened23]

Yes, i've noted this as well. I also think the prior thread was rightfully locked, the comparison was borderline troll-ish. But the point is valid: Going from Win 7 to Win 8.x with Nv resulted in an absolutely huge performance increase in BF4 using the current beta. I just play games and am not OCD about getting framerates (although I usually have MSI afterburner with the FPS counter in a corner of the screen) but performance was definitely higher under Win 8.1 with the latest drivers.

It could conceivably be the case as well for AMD drivers, but with NV's drivers the difference is very pronounced. Am curious as to whether AMD users noticed the same.

 

[itsmydamnation]

Yes I noticed this too, the recent [H] GHz review in 8.1 shows it dominating the 290 under DX. Then, within the same week Pclab runs the GHz under 8.1 and get very similar results:

Are there any other games that exhibit this behavior?

the 290 ( dont confuse the X) was always behind regular 780 in BF4, let alone a ghz , see here for a 290x to a 780.http://www.hardocp.com/article/2013...ows_7_vs_81_performance_review/4#.Uu-DufmSw7s

What about it? The beta also gained performance increases in 8.1.

The results in the second graph are from only 2 days ago.

because a lot of the performance difference was resolved. also on the other graphs are the multiplayer or single because everyone now that multiplayer is far more CPU bound then single.

BF4 has always been friendly to NV, the question is why? its unlikely driver overhead otherwise SLI would been in a far stronger position then non mantle AMD. NV dont talk that much about there internal SMX architecture ( nowhere near as much as AMD about GCN) so who knows.


edit:

looking at all the graphs, not just one (agenda much?)
http://pclab.pl/art55953-3.html

you can see AMD is bottlenecked by something, there no difference between a i3-4330 and a 4770k @ 4.5 with mantle...... AMD is leaving performance behind somewhere.......

 

[BrightCandle]

There is every chance that what happened was that in the process of adding Mantle support Dice or the frostbite guys fixed some problems. They were likely focussing on Mantle issues or things Mantle obviously exposed and in the process it helped all the cards. These issues are associated with multithreaded rendering support and because of their AMD focus I suspect they just never noticed that Nvidia gained dramatically as well due to their more parallel drivers. Interesting to see.

 

[desprado]

Can some tell that does performance increase from Win 8 to 8.1 in BF4

 

[notty22]

As far as Nvidia hardware, BF4 is not considered to be back in beta, you had to accept the patch over Origin from what I understand, is this correct ? Don't play BF4

edit, make it clear above. I don't play BF4, I didn't mean don't play BF4.

 

[itsmydamnation]

There is every chance that what happened was that in the process of adding Mantle support Dice or the frostbite guys fixed some problems. They were likely focussing on Mantle issues or things Mantle obviously exposed and in the process it helped all the cards. These issues are associated with multithreaded rendering support and because of their AMD focus I suspect they just never noticed that Nvidia gained dramatically as well due to their more parallel drivers. Interesting to see.

unlikely look at the graphs:

http://pclab.pl/art55953-3.html

if it was multithreading NV should be doing better on a 8350, what it looks like is NV is completely CPU bound where as AMD isn't with mantel. as i said before for some reason NV GPU's are performing much better on BF4 the AMD.

 

[csbin]

win8.1

Warning issued for thread crapping.
-- stahlhart

 

[csbin]

Infraction issued for thread crapping.
-- stahlhart

 

[caswow]

only pclab.pl shows 780 beeing 30% faster than 290x. nope i dont trust this

 

[omeds]

win8.1

That testing was done under win7.

Seems a few people think ArmA3 is showing improvements under win8.1 too, but I cant be certain if that's true.

 

[omeds]

because a lot of the performance difference was resolved. also on the other graphs are the multiplayer or single because everyone now that multiplayer is far more CPU bound then single.

So are you saying there's no difference between win7 and win8.1 performance? I don't follow.

All those benches are MP.

 

[Gloomy]

I'm going to wait for more Nvidia reviews that include the recent BF4 patch + Windows 8.1. If that website is reliable though... my God. 100 FPS is insane. The fact that it's bottlenecked with a stock i7-4770K is also insane. I wonder if it'll get higher at 4.8GHz?

Talking about the 780Ghz btw.

 

[Silverforce11]

So are you saying there's no difference between win7 and win8.1 performance? I don't follow.

All those benches are MP.

You have to be careful, when a lot of sites say BF4 MP, they mean empty 0/4 servers where they can run around by themselves.

You have to specifically look at 64 players listed in their review method, or they specifically say they enter a full server. Thats what computerbase.de does, ~60 players and thats what [H] does.

 

[Grooveriding]

Old news:
http://www.hardocp.com/article/2013...ows_7_vs_81_performance_review/4#.Uu99s7SwRrk


results when not in Beta:

^this

If you are actually a BF4 player you'd know the game was a mess in beta and for a short while after release with performance differences between Win 7 and Win 8/8.1. AMD and nvidia cards both improved a lot from moving to Win 8.

DICE employee on the issue

http://www.reddit.com/r/battlefield_4/comments/1r25a5/another_windows_8_upgrade_huge_increase_in/

That was fixed long ago and the two OSes perform about the same now. You should realize this because you're linking to beta benchmarks and an old bench from last year....

 

[omeds]

You have to be careful, when a lot of sites say BF4 MP, they mean empty 0/4 servers where they can run around by themselves.

You have to specifically look at 64 players listed in their review method, or they specifically say they enter a full server. Thats what computerbase.de does, ~60 players and thats what [H] does.

I'm aware of that. The problem is computerbase.de, whos reviews I regularly read and trust, tested under win7 in this instance.

 

[omeds]

That was fixed long ago and the two OSes perform about the same now. You should realize this because you're linking to beta benchmarks and an old bench from last year....

Some of those benches are under a week old and still show similar improvements on Nv hardware. That win7 vs 8 [H] review is from November and at rather GPU limited settings. Their recent GHz review is showing larger gains than expected against a 290 imho.

 

[blackened23]

You have to be careful, when a lot of sites say BF4 MP, they mean empty 0/4 servers where they can run around by themselves.

You have to specifically look at 64 players listed in their review method, or they specifically say they enter a full server. Thats what computerbase.de does, ~60 players and thats what [H] does.

PCLab also notes they entered a 64 player server. If your implication is that they entered an empty server, their testing methodology does not state that. My translation isn't perfect but the wording to me indicates a full 60+ player game as tested by PCLab, some AMD users were spreading misinformation about this to no end. As far as the other benchmark from PCGE, I do not know what their testing method for MP is.

The main point is a valid one however. NV's driver seems more fine tuned for Win 8.1 in BF4 MP, and that could well be the case with AMD as well. I don't know. AMD users probably also get a similar increase, but I don't know if the reasoning has to do with the Win 8.1 driver, or nv multithreading, or what.

 

[Silverforce11]

If anything, the BF3/4 series have clearly shown slack reviews that test SP or "simulated MP" on empty servers are meaningless since there is such a large difference when it comes to the actual intended gameplay, 64 MP.

Sites like TPU, Guru3d and AT have a lot of fame as well as resources, there's no excuse for them not testing this title the way gamers who buy it would play it.