The AMD Mantle Thread

[frozentundra123456]

I wonder why we have not seen results for 83xx cpus? Lots of slow cores would seem the ideal situation for mantle to show good results.

 

[SiliconWars]

I would wait for more results from sites like Anand's. I am skeptical of those results. That is double the maximum improvement claimed by AMD themselves as a Max case. I am not saying it is impossible, but it needs to be verified.

AMD had no way to test multiplayer, where the big gains were always likely to come.

I can't quite figure out if some people are being deliberately obtuse over this or just not seeing the simple facts.

At 1080p and below Mantle is going to be mind-blowing good - basically in any game. The main reason the gains fall off a cliff at 1600p and above is due to there being hardly any of the GPU optimizations done so far.

Everything you are seeing from BF4 Mantle is down to the basic CPU gains.

http://forum.beyond3d.com/showpost.php?p=1824246&postcount=956

That said, the gains are already there - we are seeing 10% or better gains in pathologic GPU bound cases one an inital Beta release API the first release of port of a game that was not originally designed for Mantle, in CPU bound cases the scaling is even larger; this is with everything in its infancy. As we see maturity we'll see developers going from ticking off their big ticket CPU limitation items and using Mantle features more that can bring more scaling in GPU bound cases as well.

The bolded parts are what matters.

Btw whoever asked (was it cloudfire777?) about what will happen with faster CPU's next year...the real question is what will happen with even faster GPUs, considering GPU increases at 10x the rate of CPU? Mantle is just going to improve with every year.

 

[cytg111]

rofl there are already so many tests out there its like you dont want to believe

Caution dude, caution.

edit, its kind of funny and sad, a single API vs 4 iterations of bulldozer.. wee vs meh.. AMD could have saved the money on bullfail, stayed on phenom II and done mantle to begin with. Wonder how much R&D money they coulda have saved.

 

[Despoiler]

It's like you won't provide a link to a site. AT hasn't done a test, Guru3D hasn't, HardOCP hasn't yet.

So all we have is AMD and another german site I'm not too familiar with.

Guru3d's preview

http://www.guru3d.com/articles_pages/amd_mantle_preview,1.html

 

[krumme]

That is exactly my point. Most of them show a much smaller improvement. Even amds own tests show a max improvement of half that.

LOL. Yeaa in single player bf that nobody plays. We have a perfectly documented mp test for Siege of Shanhai shown exactly where they tested - the route and the test setup. The result is not going to change. And more irrelevent sp test is not going to change that.

Everyone playing bf4 knows what this test is about, how it is to play there, and know what those results mean. Mantle dont just win. It kills DX.

 

[cytg111]

It kills DX.

- Common, you dont want to announce the death of the King before you're bonified sure you've got one to replace him.

 

[SiliconWars]

LOL. Yeaa in single player bf that nobody plays. We have a perfectly documented mp test for Siege of Shanhai shown exactly where they tested - the route and the test setup. The result is not going to change. And more irrelevent sp test is not going to change that.

Everyone playing bf4 knows what this test is about, how it is to play there, and know what those results mean. Mantle dont just win. It kills DX.

And this is without seeing Crossfire MP results. Even at 1600p Crossfire 290X should completely crush 780 SLI (1080p will be hilariously one-sided even if not particularly realistic) - and I'm just talking raw fps never-mind the clearly superior frame-times, minimums and smoothness that some of us said would be a huge benefit feature of Mantle before we even saw the benches.

 

[Erenhardt]

Mid range gpu-bound system:

Benchmark Results: In Battlefield 4, using DICE’s new benchmarking tool we received a 15.8% performance boost once Mantle was enabled at a screen resolution of 1920×1080. This is not a bad performance uplift and we’d consider a system with an Intel Core i5-3570 processor along with a Radeon R7 260X a pretty good real world platform for gamers.
Read more at http://www.legitreviews.com/amd-man...nce-catalyst-141_134959/3#5guzVthWB6YkK93Y.99

 

[desprado]

As i doubt DICE benchmark that said CF had 58% performance increase is turly for AMD marketing.
http://www.hardwareluxx.de/index.ph...te-ergebnisse-im-kampf-mantle-vs-directx.html

There really not much difference in CF.

 

[cytg111]

15% not bad.
Mantle vs directx, it does pose the question, what kind of system should I be building? Depending on what API my games use, mantle/dx, the threshold between CPU and GPU varies.. for mantle games I would shift money towards GPU power, for dx games CPU power. Glad I am not building a gaming rig right now, it really depends on mantle taking off or not.

 

[Slomo4shO]

Any Min fps numbers?

http://www.golem.de/news/amds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-3.html

Siege Of Shanghai (720p)

Direct3D 11.1: 65.4 avg / 52 min
Mantle : 116.1 avg / 88 min
​

Siege Of Shanghai (1080p)

Direct3D 11.1: 63.8 avg / 49 min
Mantle: 112.2 avg / 81 min
​

Siege Of Shanghai (1080p, 4x MSAA)

Direct3D 11.1: 57.7 avg / 44 min
Mantle: 78.2 avg / 63 min
​

Siege Of Shanghai (2160p)

Direct3D 11.1: 48.9 avg / 38 min
Mantle: 52.2 avg / 41 min​

 

[krumme]

- Common, you dont want to announce the death of the King before you're bonified sure you've got one to replace him.

Haha. No we all know dx is going nowhere

But for BF4 dx is dead. As simple as that. Mantle is so much faster it gives mantle gamers a compettitive advantage because their min fps stays up. Now go think about the consequences of that.

 

[Shivansps]

NV needs something like this, also because its important for the arm market. Mantle is OS free.

The talk that 20% does not matters doesnt make financial sense. Go and look at the price difference between cards within each brand with 20% perf difference. Its a hell of a lot of difference in revenue or reduced cost. Besides we both know the difference we see here between dx and mantle is only going to grow in time, because mantle is so new. When the gpu optimizations starts to kick in, we will have another small jump in difference.

Industry cares way more about low end and mainstream sector, 290x performance gains are bit of irrelevant for gaming industry, maybe not as much as 290X CF, but it still is.

But lets forget about all that for a moment, if mantle success, what do you think Nvidia and Intel will do? do you really belive they gona implement mantle?

Going for another solution does not help gamers, at all. Mantle is already not a good solution by itselft for AMD own cards, and im not talking about performance numbers here.

 

[desprado]

Haha. No we all know dx is going nowhere

But for BF4 dx is dead. As simple as that. Mantle is so much faster it gives mantle gamers a compettitive advantage because their min fps stays up. Now go think about the consequences of that.

As simple as that question where did u get that logic kind of u explain in detail.

 

[cytg111]

Haha. No we all know dx is going nowhere

But for BF4 dx is dead. As simple as that. Mantle is so much faster it gives mantle gamers a compettitive advantage because their min fps stays up. Now go think about the consequences of that.

Hey, i used to be a fps competitor, i know what fps'es mean. I know it sounds good, but we really dont have the complete story yet, not until the numbers have been verified, verified and then verified again. Then Ill join the quire as well.

 

[OCGuy]

Haha. No we all know dx is going nowhere

But for BF4 dx is dead. As simple as that. Mantle is so much faster it gives mantle gamers a compettitive advantage because their min fps stays up. Now go think about the consequences of that.

.....

 

[cytg111]

As simple as that question where did u get that logic kind of u explain in detail.

Imagine a good old western standoff, draw first, kill first. Simple as that.

 

[Gloomy]

AMD has BF4 and next gen non-blurry motion blur.

DX = ded

 

[Slomo4shO]

Mid range gpu

I wouldn't exactly call the R7 260X mid range, it's counterpart, the HD 7790, sells for $100 pretty regularly. Would like to see some results with the R9 270X or HD 7870.

But for BF4 dx is dead. As simple as that. Mantle is so much faster it gives mantle gamers a compettitive advantage because their min fps stays up. Now go think about the consequences of that.

Considering that Nvidia has a ~60% market share in discrete graphics...

 

[Grooveriding]

And spending hundreds of dollars more to avoid being CPU bound is the better solution for a gamer?

Mid range gpu-bound system:

This does not surprise. I think i5 systems stand to see the best gains, since they are lacking in BF4 multiplayer. I would of preferred to see at least a 270 used though.

 

[DamnedLife]

That is exactly my point. Most of them show a much smaller improvement. Even amds own tests show a max improvement of half that.

Samll improvements are all SP,
and in MP large improvements, but why still different? Because MP that's why, you can't control what other people will do..

 

[Imouto]

Siege Of Shanghai (720p)

Direct3D 11.1: 65.4 avg / 52 min
Mantle : 116.1 avg / 88 min
+77% avg / +69% min
​

Siege Of Shanghai (1080p)

Direct3D 11.1: 63.8 avg / 49 min
Mantle: 112.2 avg / 81 min
+75% avg / +65% min
​

Siege Of Shanghai (1080p, 4x MSAA)

Direct3D 11.1: 57.7 avg / 44 min
Mantle: 78.2 avg / 63 min
+35% avg / +43% min
​

Siege Of Shanghai (2160p)

Direct3D 11.1: 48.9 avg / 38 min
Mantle: 52.2 avg / 41 min
+6% avg / +8% min
​

The benefits in min framerates are quite obvious.

 

[Arkadrel]

Multiplayer results are the wow factor:
(because of more CPU stress in these situations)


From: http://www.golem.de/news/amds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-3.html

Core i7-3770K + 290x (tiny OC):

1920x1080 (no aa)
63,8 fps (dx11) vs 112,2 fps (mantle) (~75% perf. gain)




On smaller CPUs:

A10-7850k + 290:

DX11 = 31,9 fps
Mantle = 58,9 fps (~85% perf gain)

i3-4330 + 290:

dx11 = 43,3 fps
mantle = 79,7 fps (~84% perf gain)

Huge differnces in performance in multiplayer mode.

 

[Spjut]

A new multiplayer test with a comparison with Nvidia as well
http://pclab.pl/art55953-3.html

I don't believe one should look too much into AMD's worse DirectX performance for BF4, AMD doesn't really have any incentive to make specific optimizations for BF4 DX when those cards will just use Mantle instead

 

[krumme]

The test results for MP is obviously very dependent how they test it. On siege of shanghai beeing at the water scooting is different from beeing in a building.

But gameplay is defined where you are taxed most, and testing should reflect that.