Mantle vS 780Ti in BF4

KaRLiToS · 7 Feb 2014 at 14:12

I'm so freakin tired of AMD and Nvidia, we need another competitor.

TheRealDeal · 7 Feb 2014 at 14:13

GoogalyMoogaly said:
Isn't "Wiemy, co sie Kreci!" polish for "We hate Nvidia"?

I know you were joking but thought i would find out through google translate what it meant. This is the result "We know what is spinning". Hopefully they are not talking about meatspin.

LtMatt · 7 Feb 2014 at 14:15

KaRLiToS said:
I'm so freakin tired of AMD and Nvidia, we need another competitor.

Have you tried Intel?

TheRealDeal said:
I know you were joking but thought i would find out through google translate what it meant. This is the result "We know what is spinning". Hopefully they are not talking about meatspin.

Roflcopter?

TheRealDeal · 7 Feb 2014 at 14:19

Lol if the question mark is for meatspin then its not something a man wants to see put it that way. Years back i was tricked into seeing it by a phony named link. It's a little like a rude rick roll.

humbug · 7 Feb 2014 at 14:21

LtMatt said:
Source
http://www.purepc.pl/karty_graficzn...nych_battlefield_4_mantle_vs_directx?page=0,6

Thats how it should look.

thefogo · 7 Feb 2014 at 14:28

^
In DX11 with stock 4670k:
290x - 79fps
780ti - 91 fps

Interesting.

LtMatt · 7 Feb 2014 at 14:31

thefogo said:
^
In DX11 with stock 4670k:
290x - 79fps
780ti - 91 fps

Interesting.

Even weirder when you look at the i5 overclocked has the 290X in front on DX vs 780TI. It would seem the 290X is bottlenecked somehow with the i5 at stock. Unless of course something happened during that run to lower fps.

Besty · 7 Feb 2014 at 14:35

Well yeah, if you reduce the settings, the benefits of Mantle in BF4 go up.

You increase the settings, the benefits of Mantel go down.

Switching off AA and switching HBAO off will increase the differences in the 290x's favour.

Switch the AA and HBAO back on and you are back to <8% benefit.

humbug · 7 Feb 2014 at 14:35

There is no doubt Nvidia extensions have the edge in CPU threading vs AMD in DX, but its nothing to do with Microsoft's DX thread batching, this is DICE optimising for those Nvidia extensions, they don't use DX Batching as its crap.

And Nvidia's extensions certainly are nowhere near AMD's Mantle.

Silent_Scone · 7 Feb 2014 at 14:53

humbug said:
Thats how it should look when it's working

EFA

Serious note though, those tests don't have any i7 comparisons so not sure what you're comparing matey lol. Nobody is asking for proof of what performance can look like. We've seen plenty of results in the Mantle thread which are encouraging.

Gregster · 7 Feb 2014 at 14:53

Ohh look, PCper also found massive improvements with Mantle.

Results from both the Mantle version and the DX11 version are essentially the same here, averaging about 43 FPS. That's a bit disappointing but, after the preparation from AMD about setting expectations, it more or less makes sense. This is a high-end processor in a section of the game that is very GPU limited and thus we didn't expect to see a dramatic advantage for the Mantle API. We did however expect to see SOME KIND of advantage...

The denial by the AMD die hards is too funny. Site after site come up with these results and the AMD die hards dismiss them as wrong/paid off by nVidia/nVidia fan boys. I guess I am no better, as I do the same when Mantle shows astronomical gains but frankly, I see Mantle in BF4 as nothing more than a broken mess.

Matt is over the moon with it after he has to disable one card to get it to run. Others are saying "yer but Star Swarm is excellent and it is so much faster then DX" It is fundamentally broken on DX...Wake up people for goodness sake.

humbug · 7 Feb 2014 at 14:55

Using a £500 CPU its not going to make that much difference, ^^^^ at least not in this early stage of Mantle, the Key is in Mantle a much much cheaper CPU will get that same performance, that it can't get in DX. PCper don't seem to understand that. they are missing the point entirely, and so are you, again.

LtMatt said:
Even weirder when you look at the i5 overclocked has the 290X in front on DX vs 780TI. It would seem the 290X is bottlenecked somehow with the i5 at stock. Unless of course something happened during that run to lower fps.

To add to my post above you, AMD have their own Extensions for CPU threading. i don't think they are quite as good as Nvidia's, so require more optimisation, Cite BF3 which had Nvidia with a 10% advantage for a good while, AMD eventually managed to properly optimise and then overtook Nvidia, but it took a long time for them to get to that stage.

Multi-threading is not easy, it really isn't, its probably the most difficult thing to do in programming, Nvidia seem to have the better, more consistent solution, in DirectX certainly it seems that Nvidia are more efficient at dealing with multiple threads where as AMD are more efficient at dealing with IPC.
Which is so ironic as AMD CPU's lack IPC but are great in multi-threading.

I don't think AMD are putting much effort into BF4 DX Multi-threading, as they have Mantle.

LtMatt · 7 Feb 2014 at 15:00

Gregster said:
Ohh look, PCper also found massive improvements with Mantle.

The denial by the AMD die hards is too funny. Site after site come up with these results and the AMD die hards dismiss them as wrong/paid off by nVidia/nVidia fan boys. I guess I am no better, as I do the same when Mantle shows astronomical gains but frankly, I see Mantle in BF4 as nothing more than a broken mess.

Matt is over the moon with it after he has to disable one card to get it to run. Others are saying "yer but Star Swarm is excellent and it is so much faster then DX" It is fundamentally broken on DX...Wake up people for goodness sake.

Not sure why you choose to single me out Gregorio? I've said Mantle is promising but has a long way to go and needs better drivers and they need to get crossfire support working on Windows 8 asap.

I'm not sure what's up with those results you posted but even my 290 Pro card at stock clocks manages more than 43fps at 1440P 4xAA on DX/Mantle, never mind 1080P at the same settings.

TAZOO7 · 7 Feb 2014 at 15:00

And it continues, makes me laugh why half of you by either cards, as you could have saved a packet and trolled the debate with the "my life is better than your" on a £35 tablet, well you lot enjoy your ramble, im actually of to put my PC to some good use, a 8 hour session of BF4 me thinks, il try not to let the 15FPS difference put me off my game...

KaRLiToS · 7 Feb 2014 at 15:01

Gregster said:
Ohh look, PCper also found massive improvements with Mantle.

The denial by the AMD die hards is too funny. Site after site come up with these results and the AMD die hards dismiss them as wrong/paid off by nVidia/nVidia fan boys. I guess I am no better, as I do the same when Mantle shows astronomical gains but frankly, I see Mantle in BF4 as nothing more than a broken mess.

Matt is over the moon with it after he has to disable one card to get it to run. Others are saying "yer but Star Swarm is excellent and it is so much faster then DX" It is fundamentally broken on DX...Wake up people for goodness sake.

Hahahaha, so you give credibility to PcPer.

They are the most Nvidia biased butt licker.

When a baby is born, do you always critisize him because he can't walk or talk yet or you let him mature a bit before?? Why not do the same with Mantle, let it walk first.

Rusty0611 · 7 Feb 2014 at 15:04

Sees article which do not like. Attempt to question credibility of the author without proof. Repeat.

TheRealDeal · 7 Feb 2014 at 15:05

Gregster said:
Ohh look, PCper also found massive improvements with Mantle.

The denial by the AMD die hards is too funny. Site after site come up with these results and the AMD die hards dismiss them as wrong/paid off by nVidia/nVidia fan boys. I guess I am no better, as I do the same when Mantle shows astronomical gains but frankly, I see Mantle in BF4 as nothing more than a broken mess.

Matt is over the moon with it after he has to disable one card to get it to run. Others are saying "yer but Star Swarm is excellent and it is so much faster then DX" It is fundamentally broken on DX...Wake up people for goodness sake.

So from that i gather they tested in a gpu bound bit of the game on a really highend cpu and seen no real benefit's. Is it just me or am i not surprised. It was said in bf4 that the benefits will come in cpu limited situations so it's not really surprising. Why don't they test on a more realistic system that will be more widely used and in a part of the game thats more cpu taxing. Could it be because they will see actual gains.

humbug · 7 Feb 2014 at 15:06

humbug said:
Using a £500 CPU its not going to make that much difference, ^^^^ at least not in this early stage of Mantle, the Key is in Mantle a much much cheaper CPU will get that same performance, that it can't get in DX. PCper don't seem to understand that. they are missing the point entirely, and so are you, again.

Put 3 or 4 R9 290X in there and whatch DX start to bottleneck while Mantle gets its legs with that CPU and runs away with it.

PCper. Trying to make a point that does not exist.

Martini1991 · 7 Feb 2014 at 15:09

GRAPH WARS.

Silent_Scone · 7 Feb 2014 at 15:11

Well, TR constructed a CPU limited situation RealDeal. But apparently that's wrong too.

Literally nothing more you can say, any constructive criticism by either users or sites like TR just gets discredited instantly. Debates just turn into flame wars because of folk like Humbug who deflect every single bit of information rather than take the good with the bad. Just let it be folks! Not everyone is in NV pockets

lol.

Still lol'ing at the idea of foam graph fights.