AMD Mantle Support?

[Dirk Braben]

Also Mantle is supposed to be available for Linux boxes soom(tm), which is a nonneglectible thing considering SteamBox, etc.

DX will never be fully on *nix, neither will DX12, so Mantle is a good way to extend the game scene over to *nix machines - like the SteamBox and for example the Humble Bundles.

DX12 is incompatible to DX11/10/9 just like DX11 is incompatible to DX9. Games engines have to support both DX9 and DX11. And because these are incompatible to each other most games settle for DX9 because this also runs on consoles, making porting easier.
So implementing a render engine/path for DX12 will be about the same amount of work as to implement one for Mantle.

 

[Wolfs Spirit]

direct-x 12 won't be available untill late 2015 , so don't get your hope up getting it anytime soon.
DX11.2 won't have any major improvements either

Mantle is designed for AMD cpu's , thats where you get the most speedgains
Intel may just get 5-10% speedgain and with a new processor like my I7 4770k , there won't be any gain at all.
It seems they found a way to boost AMD CPU's so they can actualy compete with intell , i looked at some benchmarks and noticed up to 40% gain with older amd processors and barely anything at all with intel.

Personally if thats the only profit.. getting older systems run faster , it's a pretty stupid move from AMD , because people will wait longer buying a new system.

 

[ICDP]

direct-x 12 won't be available untill late 2015 , so don't get your hope up getting it anytime soon.
DX11.2 won't have any major improvements either

Mantle is designed for AMD cpu's , thats where you get the most speedgains
Intel may just get 5-10% speedgain and with a new processor like my I7 4770k , there won't be any gain at all.
It seems they found a way to boost AMD CPU's so they can actualy compete with intell , i looked at some benchmarks and noticed up to 40% gain with older amd processors and barely anything at all with intel.

Personally if thats the only profit.. getting older systems run faster , it's a pretty stupid move from AMD , because people will wait longer buying a new system.

Most of your quote is true, the only parts I would debate is the lack of improvement on Intel and the idea that aiming for lower end systems is a stupid move. I have a 4770K and a 290X and saw a definite improvement with Mantle over DX. In BF4 MP I was getting ~20% gains on average FPS and ~40% gains on minimums. The experience is much better with smoother FPS variance and much tighter frame pacing.

In Thief my average FPS actually dropped ~8% from 77 to 70 FPS but the minimums rose from 35 to 51 FPS. So even though the average is lower the experience is much smoother because the frame pacing is much tighter.

No doubt it does help AMD CPUs considerably, but even high(ish) end Intel CPUs such as 4770K (clocked at 4.5GHz) show definite gains.

If in the future, Mantle lowers the minimum requirements it opens up the potential for millions of people to game on a PC/laptop/tablet when they couldn't previously. AMD GCN based APUs are aimed at lower end tablets and PCs. They are Mantle compatible and if they can potentially run a AAA title even at low settings then that is a big market.

 

[Harkole]

direct-x 12 won't be available untill late 2015 , so don't get your hope up getting it anytime soon.
DX11.2 won't have any major improvements either

Mantle is designed for AMD cpu's , thats where you get the most speedgains
Intel may just get 5-10% speedgain and with a new processor like my I7 4770k , there won't be any gain at all.
It seems they found a way to boost AMD CPU's so they can actualy compete with intell , i looked at some benchmarks and noticed up to 40% gain with older amd processors and barely anything at all with intel.

Personally if thats the only profit.. getting older systems run faster , it's a pretty stupid move from AMD , because people will wait longer buying a new system.

I have a first Gen i7, just installed an R9 290x and the difference when using mantle compared to DirectX is staggering.

Tested it using the Star Scatter demo on Steam, i7 is running at native speed (3.1Ghz) with 8GB Ram. 5000 ish objects had a avg. frame rate under DX of just 23fps.

Running the same test using mantle produced fps of solid 60 (v-sync on so can't go higher).

Is it worth it, who knows. At this time the games that use it are as experimental as the API it's self. That said I do get better FPS in B4 than my Nvidia friends (similar spec PC's, but not identical my machine being about 6 months older than theirs with exception of the GPU), but they have all said they'd rather wait for DX12 than go AMD.

Personally I've had both sets of cards over the years, and for me it is as always a cost issue as the performance issue in one game is usually ruled out in another. Paying an additional £200 for the equivalent Nvidia card was a straight up no for me and so far I've not regretted the R9 at all.

 

[Im_Lost]

Also Mantle is supposed to be available for Linux boxes soom(tm), which is a nonneglectible thing considering SteamBox, etc.

DX will never be fully on *nix, neither will DX12, so Mantle is a good way to extend the game scene over to *nix machines - like the SteamBox and for example the Humble Bundles.

DX12 is incompatible to DX11/10/9 just like DX11 is incompatible to DX9. Games engines have to support both DX9 and DX11. And because these are incompatible to each other most games settle for DX9 because this also runs on consoles, making porting easier.
So implementing a render engine/path for DX12 will be about the same amount of work as to implement one for Mantle.

Yea but still, you might as well just use OpenGL instead. It's a proven API and you don't need an AMD card to run it.

 

[Robert Maynard]

Yea but still, you might as well just use OpenGL instead. It's a proven API and you don't need an AMD card to run it.

Correct me if I'm wrong, is OpenGL not rather old and limited in innovation by the "corporate" inertia of the consortium that runs it? NVidia have been offered access to Mantle, however they choose not to join in. DX12 is quite some way off - although the release of Mantle may just give MS the kick they needed to improve the release date.

 

[Malicar]

The thought of running Mantle on an engine that truly supports multiple cores such as the Nitrous engine is mind blowing. Then thinking about perhaps unified memory designs with GDDR5 and possible boards with flash ROM OS memory to do OS installs online and a bunch of other next gen advancements make Malicar happy

While I'm a Intel\Nvidia user for many years now it does appear AMD is fighting back. Their new GPU is smoking awesome and I really have to say for the first time in years I was on the AMD site looking at 8 core CPU's. Also AMD is doing well with their APU's and I don't see them going any time soon. I have to admit I've considered building an AMD gaming PC.

Plus when you look at where Intel is headed there are really no big gains in the CPU market looking ahead. Unless your into Intel HD graphics integration. Most of us will be scratching our heads wondering where the new faster CPU's are. So does this mean in order to push the limits we are now going to have to shift into true multicore engines and take AMD's approach with more physical cores and proper thread usaging across them all?

Apparently the Nitrous engine also enhances Haswell support... Either way Nitrous has my attention and so does Mantle.

 

[Bolthammer]

As much as i like the underdog AMD cpu wise, and owned my fair share of ati cards and i like improvements. I respect you wuld like all developers to embrace mantle because you own a Mantle-able card.
But then we culd make huge posts about how to embrace Intels hyperthrading more, Nvidias physx or gsync, xbox kinect etc.
Im sure FD will implement it if its worthwile for their customers. If they deem the improvements low priority since the game isnt that gpu-taxing at the moment, they might implement them down the line. As i read it, DX12 will support Mantle in some way or form also.

Its not that important, rather have stable netcode, feature rich games, new content etc. Half the gpu optimization for games happens from amd and nvidia themselves if they deem the game important fx benchmarking/reviews.

So lets get a solid Elite demo/benchmark that is used by reviewers. That will certainly see some improvements hidden in the drivers.

 

[Mazhurg]

What many of the posters on Mantle here fail to understand, is that the requirements for it was generated by developers, not AMD.

AMD provided Mantle upon their requests, and is one of the main reasons why it is being adopted by a multitude of game engines companies.

 

[Im_Lost]

Correct me if I'm wrong, is OpenGL not rather old and limited in innovation by the "corporate" inertia of the consortium that runs it? NVidia have been offered access to Mantle, however they choose not to join in. DX12 is quite some way off - although the release of Mantle may just give MS the kick they needed to improve the release date.

OpenGL can do pretty much all the things DX can, sometimes slightly better. The PS4 uses it and Linux is still becoming more popular so it is being kept up with the times. It's also open source unlike DX, which is a huge plus.

 

[HaiderGill]

direct-x 12 won't be available untill late 2015 , so don't get your hope up getting it anytime soon.
DX11.2 won't have any major improvements either

Mantle is designed for AMD cpu's , thats where you get the most speedgains
Intel may just get 5-10% speedgain and with a new processor like my I7 4770k , there won't be any gain at all.
It seems they found a way to boost AMD CPU's so they can actualy compete with intell , i looked at some benchmarks and noticed up to 40% gain with older amd processors and barely anything at all with intel.

Personally if thats the only profit.. getting older systems run faster , it's a pretty stupid move from AMD , because people will wait longer buying a new system.

Draw calls are a lot faster on consoles than PCs. Mantle will help even this out regardless of Intel or AMD CPUs. As people get more experienced with Mantle there will be more and bigger gains. People are still learning to use it...

 

[Bolthammer]

In time it might cure diseases and stop asteroids from hitting earth to.

It wont be the next great thing unless it easily incorporated in all existing gpu-vendors both software and silicon-side. If not - then it will only be a amd gimmick like physx. Nice to have but not need to have.

 

[HaiderGill]

In time it might cure diseases and stop asteroids from hitting earth to.

It wont be the next great thing unless it easily incorporated in all existing gpu-vendors both software and silicon-side. If not - then it will only be a amd gimmick like physx. Nice to have but not need to have.

It all depends really. It's not meant to be used by most developers just the AAA software houses.

 

[HaiderGill]

Hopefully Mantle will be incorporated, but G-sync is just dependant on GPU drivers I think, it isn't game specific as far as I recall.

Gync is good. I read this regarding AMD's potential implementation http://www.extremetech.com/gaming/1...with-a-free-driver-update-nvidia-rebuts-fight. Have to see how it work in the real-world.

 

[Bolthammer]

But then again shuld gpu-engineers make cores on a gpu to suit Mantle? Seems bad to to handcuff development to a certain way of thinking.

I cant see Nvidia interested in adopting Mantle at all. They are working hard on putting arm-cores into their gpus. They think it will give same or better benefit than Mantle. They will use the arm-core for frame and driver calls and in return hope for better latency and response times.

What is best, smartest or most brutal way to get the same results dossent matter. Its how they execute their pr, implementation, powerefficiency and how little work developers have to put in for sexy frames. And price!!!

We all remember how it went with 3Dfx and Matrox, they were both held high in regard to their forward thinking or 2d quality. In the end Nvidia thrumped them with brute force, then came Ati with some good brutal series.

Now the battle is between efficient and smart Mantle but less powerefficient top end cards vs somewhat powerefficient cards nvidia cards with their Maxwell series promissing same efficiency as current series but with more grunt. Wich PR machine will be best, wich will be the benchmark king and most importantly wich vendor makes the cheapest and best card in the gtx770/780/280/290 pricerange. Mantle, gsync or insert name of smart thinking matters less.

 

[Arowx]

It sounds like OpenGL and DirectX 12 are now addressing the bottlenecks that Mantle addresses as well as adding Mantle style features.

So Mantle has done a good job of shaking up the market a bit, which is great for the PC platform and future mobile devices.

http://www.tomshardware.com/news/directx-direct3d-opengl-mantle,26167.html

http://www.pcper.com/news/Graphics-Cards/DirectX-12-and-new-OpenGL-challenge-AMD-Mantle-coming-GDC

 

[Malicar]

It sounds like OpenGL and DirectX 12 are now addressing the bottlenecks that Mantle addresses as well as adding Mantle style features.

So Mantle has done a good job of shaking up the market a bit, which is great for the PC platform and future mobile devices.

http://www.tomshardware.com/news/directx-direct3d-opengl-mantle,26167.html

Even better for an 8 core CPU like the AMD ones which pound for pound are slower than their intel brethren but have more cores. I'm also excited about Mantle. Nitrous engine has me curious too. I think even Stardock is using it for Galciv3 if I'm not mistaken. In theory PS4 could also use it. Microsoft probably wouldn't because of DX12 but Sony could reap best of both worlds and maximize developer bias.

 

[HaiderGill]

But then again shuld gpu-engineers make cores on a gpu to suit Mantle? Seems bad to to handcuff development to a certain way of thinking.

I cant see Nvidia interested in adopting Mantle at all. They are working hard on putting arm-cores into their gpus. They think it will give same or better benefit than Mantle. They will use the arm-core for frame and driver calls and in return hope for better latency and response times.

What is best, smartest or most brutal way to get the same results dossent matter. Its how they execute their pr, implementation, powerefficiency and how little work developers have to put in for sexy frames. And price!!!

We all remember how it went with 3Dfx and Matrox, they were both held high in regard to their forward thinking or 2d quality. In the end Nvidia thrumped them with brute force, then came Ati with some good brutal series.

Now the battle is between efficient and smart Mantle but less powerefficient top end cards vs somewhat powerefficient cards nvidia cards with their Maxwell series promissing same efficiency as current series but with more grunt. Wich PR machine will be best, wich will be the benchmark king and most importantly wich vendor makes the cheapest and best card in the gtx770/780/280/290 pricerange. Mantle, gsync or insert name of smart thinking matters less.

Brute force equasl heat and subsequently fan noise...

 

[HalloweenJack]

heya all

ccame here from google - was looking about ED and mantle - seems a few misconceptions in here about it though


if you read some reviews on anad , hex , tpu etc - on mantle enabled games , you`ll find the greatest bebefit is for low end machines - eg APU based, which can move more of the load from the weaker CPU to the better GPU.


This IMO would be of great benefit for ED.

why?

it lowers the minimum spec - OR , gives the minimum spec better playability

AMD A6>A10 APU (maybe even with hybrid) running mantle? easier porting to consoles? Intel themselves are looking at mantle - as it would benefit those i3`s using ondie gpu`s - in those £399 pc world boxes.

whilst its great so many have high end rigs costing more than a cheap car , a lot of families cant throw £800+ at a pc , so IMO using mantle would open up ED to an even better audience

 

[Hawkmoon]

heya all

ccame here from google - was looking about ED and mantle - seems a few misconceptions in here about it though


if you read some reviews on anad , hex , tpu etc - on mantle enabled games , you`ll find the greatest bebefit is for low end machines - eg APU based, which can move more of the load from the weaker CPU to the better GPU.


This IMO would be of great benefit for ED.

why?

it lowers the minimum spec - OR , gives the minimum spec better playability

AMD A6>A10 APU (maybe even with hybrid) running mantle? easier porting to consoles? Intel themselves are looking at mantle - as it would benefit those i3`s using ondie gpu`s - in those £399 pc world boxes.

whilst its great so many have high end rigs costing more than a cheap car , a lot of families cant throw £800+ at a pc , so IMO using mantle would open up ED to an even better audience

Totally agree with you, if you look at the results of tests the biggest increase in performance is on lower end systems, and with some of the most resource hungry games and apps (BF4, Cryengine) supporting it you can see that it will boost sales of APU based systems for gaming. I don't completely agree with people saying it, that developers will not support it if it is hardware specific, I can see where they are coming from, but you also have to take into account how many more systems will be able to run a new title at a usable frame rate, an extra 10-20% sales would make the extra work well worth it. For FD I think it is a case of Mantle coming on the scene a little too late, if it had been launches about eighteen months earlier (so was where it is now when the Devs really got to work) then I think we might well have seen support for it in ED.