Mantle API presentation by AMD, DICE and Oxide - AMD Summit 2013

[erocker]

While AMD is allegedly releasing a new driver (it has a news post and everything) there is no sign of Mantle or TrueAudio in there.

No. That driver was from Toshiba and has nothing to do with the next driver. The driver itself is older than the latest beta on AMD's site.

 

[xenocide]

No. That driver was from Toshiba and has nothing to do with the next driver. The driver itself is older than the latest beta on AMD's site.

Ah my mistake. I apparently skimmed over that part.

 

[natr0n]

Everything happens today.

 

[broken pixel]

@ 4:00am EST you get the BF4 Mantle patch, the AMD Mantle driver is?

 

[BiggieShady]

http://www.pcper.com/reviews/Graphi...rings-Mantle-Support-Frame-Pacing-Phase-2-HSA

So Mantle is released as an early beta with this driver ... and cross fire frame pacing ... and HSA support ... well done AMD *clap* *clap*

 

[the54thvoid]

It'll be good to see the actual FPS graphs from those %'s. For those on APU's and high end GFX it'll be a massive boost but I'd like to know what the FPS are in the first instance. How much oomph do you get with a Kaveri and 290X boosted with Mantle, versus (for example) a 4930k and 290X or more flamebait worthy, a 4930k and a 780ti? I'm on a 3930k at 4.4GHz so I expect I'd see very low returns with Mantle (if I had a GCN card).
What we all want to see (or some really don't want after all the trash talk ) is with a powerfully specced CPU, will Mantle make a significant impact on fps in a 290x versus 780ti scenario?

And FTR, where's MSI's 290X Lightning at?

 

[KingPing]

The Star swarm benchmark is out now on Steam, it's Dx11 capable too.

 

[BiggieShady]

with a powerfully specced CPU, will Mantle make a significant impact on fps in a 290x versus 780ti scenario?

Define significant ... article mentions 10% frame rate increase on high end cpus, and 40% frame rate increase on kaveri in BF4. With multi gpu setups where you are rarely gpu bound, it should be similar increase for all kinds of cpus.

 

[TheHunter]

The Star swarm benchmark is out now on Steam, it's Dx11 capable too.

Yep, so far it looks like DirectX11 is limited to avg ~ 3800 units, doesnt matter which gpu.


With Mantle +10000 units

 

[crazyeyesreaper]

So essentially on really low end systems which no one around here really runs theres a huge performance gain. However on hardware thats of the high end or enthusiast class theres about a 10% gain?...... I could have sworn i said maybe 20% at best..... and here it seems to be 10% which can be offset with a GPU overclock. Mantle does not seem that impressive. Then again BF4 is not the best choice for Mantle anyway. an RTS would have made more sense. The Star Swarm benchmark is interesting but its not a game. Lets see Star Craft / Total War / Company of Heroes / World in Conflict / etc etc etc with Mantle. the benefit in those games would actually be noticeable.

 

[TheHunter]

Any kind of boost is welcome imo even if its minimal.



And from what I saw 7970 can have avg 25% boost..

 

[crazyeyesreaper]

25% on an AMD CPU

on an Intel CPU it seems to be 7.5% to 10%

it just puts AMD and Intel on equal footing but does not impact playability that much. We have seen larger performance increases from simple VGA driver updates lol.

• Core i7-4960X CPU + R9 290X GPU
  • 1080p, Ultra Preset, 4xAA: 9.2% improvement with Mantle
  • 1600p, Ultra Preset, 4xAA: 10% improvement with Mantle
• Core i7-4960X CPU + R7 260X GPU
  • 1080p, Ultra Preset, 4xAA: 2.7% improvement
  • 1600p, Ultra Preset, 4xAA: 1.4% improvement
• A10-7700K CPU + R9 290X GPU
  • 1080p, Ultra Preset, 4xAA: 40.9% improvement
  • 1600p, Ultra Preset, 4xAA: 17.3% improvement
• A10-7700K CPU + R7 260X GPU
  • 1080p, Ultra Preset, 4xAA: 8.3% improvement
  • 1600p, Low Preset: 16.8% improvement

So from what I can see here is that the performance change from Mantle on a high end system are less then what can be achieved through overclocks and traditional driver optimizations.

Any kind of boost is welcome imo even if its minimal.



And from what I saw 7970 can have avg 25% boost..

Low end system getting such ahuge boost is nice but then how many people here are gonna buy a A10 7700k and pair it with a $600 GPU. Even you have a 4770K so at best if you got a 290X with 10% boost it would offer enough performance to be on par with a 780Ti in BF4.

so..... yea not really that impressive. 290X ends up out pacing a stock 780Ti but aftermarket 780 Ti beats the aftermarket 290X even with Mantle. At worst case senario they Tie. Yes AMD is cheaper but I would still take the Nvidia card at this point.

 

[Crap Daddy]

Any kind of boost is welcome imo even if its minimal.



And from what I saw 7970 can have avg 25% boost..

With an AMD CPU.

 

[BiggieShady]

BF4 is not the best choice for Mantle anyway. an RTS would have made more sense. The Star Swarm benchmark is interesting but its not a game. Lets see Star Craft / Total War / Company of Heroes / World in Conflict / etc etc etc with Mantle. the benefit in those games would actually be noticeable.

Exactly. All that Mantle does now is open the possibilities for developers. We can't expect to see full benefits until whole new generation of games gets designed with wider constraints in mind. Even then it will be less performance benefits and more eye candy ... there will be situation where game's DX11 mode looks worse because, for example, "very high" setting for view distance in DX11 mode is 2 km and in mantle mode it's 10 km.

 

[crazyeyesreaper]

Thats all well and good and I have said that all along but the trolls here including the one who started the thread like to continually tell me I am wrong and have no damn idea whats going on. Its honestly a joke we have laughed about in the TPU teamspeak for awhile now.

Mantle wont change anything anytime soon yet the believers are as bad as the beliebers in their unflinching devotion that its inclusion in BF4 means instant success.

Lets face it A10 7700k + 290x works well in BF4
now lets take a look at the entire Steam / Origin / GOG library and realize its a drop in the ocean right now. It has no meaning Mantle wont matter for a number of years. When it does matter the current implementation of GCN will be much like the HD 5000 series is now. Old and nearly forgotten.

 

[TheHunter]

Then you two didnt see this yet..

Test case 1: Low-end single-player
CPU/GPU: AMD A10-7850K (‘Kaveri’ APU), 4 cores @ 3.7 GHz
Settings: 720p MEDIUM settings.
OS: Windows 7 64-bit
Level: Singapore “Beach”
This area is heavy on the CPU as it is quite a lot of action going on, but the game is primarily GPU-bound both on DX and with Mantle but thanks to GPU optimizations with Mantle we get a good performance improvement.
Result: 26.6 ms/f -> 23.3 ms/f = 14% faster

Test case 2: Standard 64-player multiplayer
CPU: AMD FX-8350, 8 cores @ 4 GHz
GPU: AMD Radeon 7970 3 GB
Settings: 1080p ULTRA 1x MSAA
OS: Windows 8 64-bit
Level: Siege of Shanghai
Level was tested with 64 “pseudo players” that we have for our own internal testing that simulates heavy game workload that we have in multiplayer in order to get more deterministic results compared to full real multiplayer. 64 players on the large Battlefield levels is really demanding of the CPU so this test case is primarily CPU-bound.
Result: 18.87 ms/f -> 15.08 ms/f = 25.1% faster

Test case 3: High-end single-player with multiple GPUs
CPU: Intel Core i7-3970x Extreme, 12 logical cores @ 3.5 GHz
GPU: 2x AMD Radeon R9 290x 4 GB
Settings: 1080p ULTRA 4x MSAA
OS: Windows 8 64-bit
Level: South China Sea “Broken Flight Deck”
This single-player scene is heavy on both the CPU and GPU with lots of action going on. Test was done on the highest end Intel CPU on Windows 8, which is the fastest option before Mantle thanks to DirectX 11.1. Still this CPU is not fast enough to keep the 2 290x GPUs fed at 1080p on Ultra settings so we get a significant CPU performance bottleneck which results in major performance improvement when enabling Mantle.
Result: 13.24 ms/f -> = 8.38 ms/f = 58% faster

straight from battlelog.

 

[FordGT90Concept]

The Star Swarm benchmark is interesting but its not a game. Lets see Star Craft / Total War / Company of Heroes / World in Conflict / etc etc etc with Mantle. the benefit in those games would actually be noticeable.

Command & Conquer: Generals 2

That's the only big RTS title coming up I can think of that is likely to support Mantle.

 

[BiggieShady]

When it does matter the current implementation of GCN will be much like the HD 5000 series is now. Old and nearly forgotten.

Funny enough, AMD team addressed that exact issue when presenting Mantle saying they have targeted right level of API abstraction to allow them two things: to make it work with next versions of GCN, and to make it work with competitors GPUs. Essentially they said wait for these two things to happen

 

[crazyeyesreaper]

My point is performance wise it wont matter.

oh wow so AMD fixed Crossfire X with Mantle...... huh thats not surprising in the least. They needed to change how Crossfire worked at a base level to fix it. so of course mantle allows xfire to work better but again how many buy 2x $600-700 GPUs because righ tnow the 290X is $700 and the 290 is $500+ so still $1000-1400 and the only game that benefits is BF4 a Game that a single gpu maxes without issue at 1200p no problem. I see the benefits I just don't care they don't offer enough change to make them look better its all specific situations.

They picked BF4 because its market presence is huge. Sadly an FPS is not the best way to show off improved CPU utilization.

again Star Craft / Total War / Company of Heroes / etc are better options for showing marked gains with a new API. BF4 just wasnt that bottlenecked. The frame pacing aspect is a nice change but generally it just puts them closer to and on par with Nvidia which good but not mind blowing.

As for waiting for support from Nvidia sure we can wait not that big a deal but so far BF4 and Thief which are big mantle titles are a MESS.

At this point Thief is shaping up to be a flop. and BF4 is broken still so its a better running game thats still broken lol.

You could say AMD has continued to be AMD with a launch that offers hope but really just doesn't live up the hype train.

 

[erocker]

They picked BF4 because its market presence is huge. Sadly an FPS is not the best way to show off improved CPU utilization.

They picked BF4 because DICE said "yes".

 

[crazyeyesreaper]

That to but doesnt change the fact its a paltry 9-10% which considering BF4 performance means the 290x is now = 780Ti in BF4 and they both cost the same on Newegg. So....... not really impressive haha.

 

[erocker]

I don't care what cards cost. Those with cards already get a free boost, be it small or large to work with. It's something... and it's free.

 

[TheHunter]

Yawn..

Not worth the energy to argue with diehard nv fanboys,
laterz.

 

[Crap Daddy]

That to but doesnt change the fact its a paltry 9-10% which considering BF4 performance means the 290x is now = 780Ti in BF4 and they both cost the same on Newegg. So....... not really impressive haha.

I remember an AMD rep saying that with mantle the 290x will dwarf the Titan. How things change in a few months. Nevertheless we should wait for some real benchies. Thing is it seems it's not enough you have an AMD card, now you also need an FX!

 

[crazyeyesreaper]

Yea it is funny how things change. I also love how everyones default reply is oh your a fanboy. When sadly I have probably owned more Radeon GPUs then the people claiming I am a fanboy lol.
AMD
4870x2
5850 x2
6950 x2
6970 x2
7970

NVIDIA
7800 GTX
8800 GTS 640MB
GTX 780

lol oh well. I myself am not impressed granted I didnt believe the bs PR and expectations either. Single player gains are worthless in a game where the single player was worthless and people only played it to unlock weapons for MP. In two to three years Mantle has the ability to be amazing BF4 just doesn't benefit much from the API unless your PC is shit to begin with.