The AMD Mantle Thread

[Saylick]

Nobody that dare to give a percentage boost on Mantle? Say average boost, since some here say games will differ.

Come on, after all this praising, time to put your money where your mouth is

+5%? +10%? +30%? 10-20%?

You could always start a poll... :hmm:

On the other hand, there are far too many variables to account for. The performance increase for a mid-range model may differ from the increase for a high-end model.

 

[carage]

I'm just worried this is Glide all over again.
When I saw the Mantle presentation, the first thought that came to mind was exactly, oh s**t, 3dfx is back.
The difference between this and PhysX is that AMD now also owns the console market so they have the relevant market power to force/push/persuade developers to come onboard or else. It will be interesting if Microsoft has anything to say about this. On the PC front, this could totally screw up DirectX.

 

[Kenmitch]

Nobody that dare to give a percentage boost on Mantle? Say average boost, since some here say games will differ.

Come on, after all this praising, time to put your money where your mouth is

+5%? +10%? +30%? 10-20%?

Average when? Only statement about Mantle so far is BF4 patch in December.

<---- Closes eyes really tight, concentrates, views future benchmarks showing mantle in BF4, squints but can't make out the numbers kinda looks like a 23.4% increase....Might be a 8 tho.

I'm keeping my pot....Not sure why the other guy is willing to give £50 in pot.

I'm just worried this is Glide all over again.
When I saw the Mantle presentation, the first thought that came to mind was exactly, oh s**t, 3dfx is back.
The difference between this and PhysX is that AMD now also owns the console market so they have the relevant market power to force/push/persuade developers to come onboard or else. It will be interesting if Microsoft has anything to say about this. On the PC front, this could totally screw up DirectX.

NVidia gladly handed AMD the console market....At least that's what they said anyways.

Yes it could screw up DirectX but on the other hand it could force Microsoft to make improvements also.

 

[Cloudfire777]

You could always start a poll... :hmm:

On the other hand, there are far too many variables to account for. The performance increase for a mid-range model may differ from the increase for a high-end model.

The thing is, nobody here have remotely any idea what Mantle will do. Nobody here have any first hand experience with API coding regarding graphic cards, but somehow many seem pretty certain that it will be amazing.

Based on what....? Consoles? Then what boost do you know of that these get out of the hardware, using similar APIs, if you compare them against the same hardware running on PCs?

They dont know that either.

 

[Shivansps]

The thing is, nobody here have remotely any idea what Mantle will do. Nobody here have any first hand experience with API coding regarding graphic cards, but somehow many seem pretty certain that it will be amazing.

Based on what....? Consoles? Then what boost do you know of that these get out of the hardware, using similar APIs, if you compare them against the same hardware running on PCs?

They dont know that either.

its called marketing, launch a bomb, give minimum info as possible, start selling the things and give the full details a month after the sales, when the hype already died out.

 

[DarkKnightDude]

This part in the DICE slide was interesting...

http://www.slideshare.net/fullscreen/DICEStudio/battlefield-4-frostbite-3-mantle/20

"low-latency multi-GPU rendering for VR"

 

[Saylick]

The thing is, nobody here have remotely any idea what Mantle will do. Nobody here have any first hand experience with API coding regarding graphic cards, but somehow many seem pretty certain that it will be amazing.

Based on what....? Consoles? Then what boost do you know of that these get out of the hardware, using similar APIs, if you compare them against the same hardware running on PCs?

They dont know that either.

I know what you're saying and I agree; everything thus far about the next-gen graphics cards from AMD is rumored and unconfirmed.

The reason why people are hopeful is that Mantle, in concept, can be a game changer. Logically speaking, why would a company make a push for something like Mantle if the potential benefits didn't outweighed the risks? On the flip side, you also have people who are skeptical as a result of prior experiences with AMD promising something over the top.

As for how much will Mantle help GCN hardware, I made the same argument you just made about the benefits of Mantle a while back. Realistically, Mantle will most likely give a performance boost to PC hardware, but the million dollar question is that if hypothetically speaking a Mantle-esque, low-overhead API provides a 50% boost to console hardware vs. consoles not using said API, how much of that 50% is transferable to PC hardware? It's fair to say that the transfer efficiency will be less than 100% but how much lower? Some may say that we can expect somewhere in the region of 20%-30% improvement using Mantle, some are less optimistic and give a figure around 5%. I'm thinking it will be at least 10% but definitely not large enough that nVidia will be instantly crapping their pants, i.e. 50%.

With that said, I don't have a background in programming at all, let alone computer graphics, so my guess is probably worth zilch. The best I can do is postulate based on reasoning, logic, and viewing the situation as if I was a third-party spectator in a chess match: I have no idea what the chess players are thinking, but I can get an idea of what is going to happen and what the chess players are hoping will happen by looking at the positions of the pieces on the board. In this chess analogy, Mantle is essentially an old school strategy that hasn't been seen nor used in ages, where the outcome heavily depends on the set-up and that only a handful of people have seen said strategy in play. In this case, one can only speculate on the effectiveness of said strategy by analyzing the playing field and all prior moves in the game.

 

[Keysplayr]

A person with NVidia hardware can only hope for this to be the situation.

By the statements made about Titan being ridiculed in BF4 by 290x with Mantle make most think otherwise. Ridiculed to me doesn't sound like a couple of fps.

What's your take on Mantle? What was your response when you 1st heard of it? I'm just curious is all and I'm not trying to start any trouble FYI.

Don't hype this up too much Kenmitch (and others) as you are likely to be let down if expectations are as high as Everest. I mean, learn.
It'll be given a chance just like anything else, and I hope they can even do half as good as they are pimping.

 

[Silverforce11]

Saylick, its more like 2.5x the performance boost of console hardware versus PC going through DX, that figure is touted by big name devs as the DX overhead.

The question is how much of that 2.5x will translate to PC running GCN. I am not overly optimistic because a lot of the gains on the console is due to lifting CPU bottlenecks as well as GPU optimizations. On the PC, the CPU is rarely the bottleneck. Still, I expect it will be significant, else why would devs be interested in changing their game engine to support a new API?

 

[BoFox]

Not really. PhysX is just an added feature, I have an AMD card, and dont really miss it at all. In fact, I was running CPU physX on Borderlands 2 and turned it off because I just found all the sparkly effects annoying. Adding physX didnt allow the developer to bypass Direct X, thus not making them devote extra resources to it. The potential problem I see with Mantle is that is could divert developers from devoting sufficient resources to DirectX, thus making performance poor on any non-AMD cards.

I suppose you can roll your eyes at Anand's article as well as my feelings if you wish, but apparent it is a legitimate concern.

The up-side to this is that M$ will have more reasons to improve DirectX thanks to Mantle, and thanks to NV pushing harder for it - they already rolled out DX11.2 for XBone (and Win8.1). NV didn't bother with full DX11.1 or DX11.2 support with the Kepler series, but don't be surprised if NV goes for DX12 with Maxwell.

Don't worry - Nvidia is always known to be a fiercely competitive company, so this will only drive NV "mad" in the spirit of competition. With a superb driver team, NV will try their hardest to work closely with the same game developers who code for Mantle for the PC, to ensure that their cards also perform well. NV definitely has more resources, after all. JHH was probably regretting not cementing in a console deal after showing up at AMD's event, but his mind is probably set on the Android market - the real killer that got Qualcomm sky-high, or on being like Samsung or even Google/Apple (not just dozens of billions of dollars, but more..) after disappointing business with the consoles in the past for tiny pocket change. Only if Qualcomm were still under AMD.......

One thing I dislike though is AMD's rumored statement that Stereo3D is a failure, that AMD will be focusing on 4k displays instead of S3D. To slack off in any department can really back-fire, just like it did with Intel's abandoning of the discrete GPU market only to have to license NV's tech later on.

What if Samsung indeed buys AMD? IBM could do that as well. Apple might too, but Apple's personality is just not anything like AMD at all. But then again, Apple just might, because AMD's net worth is meager pocket change. The company that should buy AMD the most is Qualcomm, IMHO.. with net worth of $115 billion, it doesn't hurt to snatch up every asset of AMD (2.78 billion plus ~2B debt), and digest every useful bit, then create hybrid x86/ARM PC's and everything else, etc... while even making an all-in-one console in the future, GPU cloud server farms, etc.. Man, I tell you - these cloud server farms are going to be a massively huge thing in the future. Microsoft could be another potential buyer.... AMD is still a diamond in the rough - Qualcomm came out of it, and look what Qualcomm turned into after some polishing/cutting. I wish I had $5 billion to buy AMD (while having some other assets), and then drive AMD into something bigger than Intel and Qualcomm and Samsung combined.

 

[Saylick]

Saylick, its more like 2.5x the performance boost of console hardware versus PC going through DX, that figure is touted by big name devs as the DX overhead.

The question is how much of that 2.5x will translate to PC running GCN. I am not overly optimistic because a lot of the gains on the console is due to lifting CPU bottlenecks as well as GPU optimizations. On the PC, the CPU is rarely the bottleneck. Still, I expect it will be significant, else why would devs be interested in changing their game engine to support a new API?

I recall reading about figures around that mark as well, but I was just giving a hypothetical situation in my previous post. More importantly, like you stated, the biggest detail lies in how much of that performance increase is transferable to PCs.

 

[ShintaiDK]

http://vr-zone.com/articles/mantling-alliances-ritchie-corpus-amd-interview/58215.html

If you think about it, Mantle is truly a collaboration and I&#8217;ll tell you [DICE&#8217;s Johan Andersson] was at the forefront of that at the very beginning. A lot of the feedback on the development of Mantle came from him. We also solicited feedback from a lot of other partners that we haven&#8217;t announced yet. At this stage, Battlefield 4 and FrostBite 3 are the closest to deliver something today. I think, as I mentioned before, the goal would be to provide the spec and SDK publicly.

VRZ: Any timeline to Mantle&#8217;s SDK being made public?

It could be as early as sometime next year or maybe the year after.

So, DICE only. No specs or SDK until...who knows when. Not exactly something that will be adopted anytime soon. Assuming it ever will be.

VRZ: Valve has recently announced that the Steam Box is in beta. In Valve&#8217;s announcing press release, it said how Nvidia is a big stakeholder in the Steam Box. What is AMD doing to counter or compete against this?

There is no counter.

The reason is we are working just as closely with Valve. I think the difference is one side of that conversation is being more vocal than the other. If you go to Valve right now and ask them which hardware partners are going to be the partner of choice, they won&#8217;t pick one or the other. They are going to be agnostic. They mention that it&#8217;s an open ecosystem.

No Mantle for SteamOS.

 

[SiliconWars]

So, DICE only. No specs or SDK until...who knows when. Not exactly something that will be adopted anytime soon. Assuming it ever will be.

Uhh, that's nothing like what was said. Clearly there are unannounced partners - the hint being "We also solicited feedback from a lot of other partners that we haven’t announced yet."

No Mantle for SteamOS.

Was not said either. In fact I have no idea how you could conclude that based on what was said. Or well, I do have an idea.

 

[3DVagabond]

Nobody that dare to give a percentage boost on Mantle? Say average boost, since some here say games will differ.

Come on, after all this praising, time to put your money where your mouth is

+5%? +10%? +30%? 10-20%?

I'd say that would depend on how it's implemented. If you still use Dx11+ and can simply code some individual parameters with mantle it could vary wildly depending on what the overhead is on those parameters. A few percent... 20%? If though, you have to use either all Mantle or all Dx the increase using purely Mantle for everything should be dramatic. Twice as fast, or more, is possible. This will also increase as devs get better with optimizing for Mantle just like with consoles as they get better over time. Think of the difference between early Xbox games and the latest ones.

 

[3DVagabond]

I'm just worried this is Glide all over again.
When I saw the Mantle presentation, the first thought that came to mind was exactly, oh s**t, 3dfx is back.
The difference between this and PhysX is that AMD now also owns the console market so they have the relevant market power to force/push/persuade developers to come onboard or else. It will be interesting if Microsoft has anything to say about this. On the PC front, this could totally screw up DirectX.

Or else what? What do you think AMD could do to strong arm them?

Devs are going to be using it for the consoles already. Unless they just don't care how well their games run on the PC, they're going to use it. They've already coded it into the console versions of the games. Especially if it's as Ryan Smith believes and is exactly the same as Xbox1's API.

 

[3DVagabond]

The up-side to this is that M$ will have more reasons to improve DirectX thanks to Mantle, and thanks to NV pushing harder for it - they already rolled out DX11.2 for XBone (and Win8.1). NV didn't bother with full DX11.1 or DX11.2 support with the Kepler series, but don't be surprised if NV goes for DX12 with Maxwell.

Don't worry - Nvidia is always known to be a fiercely competitive company, so this will only drive NV "mad" in the spirit of competition. With a superb driver team, NV will try their hardest to work closely with the same game developers who code for Mantle for the PC, to ensure that their cards also perform well. NV definitely has more resources, after all. JHH was probably regretting not cementing in a console deal after showing up at AMD's event, but his mind is probably set on the Android market - the real killer that got Qualcomm sky-high, or on being like Samsung or even Google/Apple (not just dozens of billions of dollars, but more..) after disappointing business with the consoles in the past for tiny pocket change. Only if Qualcomm were still under AMD.......

One thing I dislike though is AMD's rumored statement that Stereo3D is a failure, that AMD will be focusing on 4k displays instead of S3D. To slack off in any department can really back-fire, just like it did with Intel's abandoning of the discrete GPU market only to have to license NV's tech later on.

What if Samsung indeed buys AMD? IBM could do that as well. Apple might too, but Apple's personality is just not anything like AMD at all. But then again, Apple just might, because AMD's net worth is meager pocket change. The company that should buy AMD the most is Qualcomm, IMHO.. with net worth of $115 billion, it doesn't hurt to snatch up every asset of AMD (2.78 billion plus ~2B debt), and digest every useful bit, then create hybrid x86/ARM PC's and everything else, etc... while even making an all-in-one console in the future, GPU cloud server farms, etc.. Man, I tell you - these cloud server farms are going to be a massively huge thing in the future. Microsoft could be another potential buyer.... AMD is still a diamond in the rough - Qualcomm came out of it, and look what Qualcomm turned into after some polishing/cutting. I wish I had $5 billion to buy AMD (while having some other assets), and then drive AMD into something bigger than Intel and Qualcomm and Samsung combined.

I don't know how many times this has been said, but if you buy AMD you don't get their X86 license.

 

[Grooveriding]

http://vr-zone.com/articles/mantling-alliances-ritchie-corpus-amd-interview/58215.html



So, DICE only. No specs or SDK until...who knows when. Not exactly something that will be adopted anytime soon. Assuming it ever will be.



No Mantle for SteamOS.

Considering Gabe Newell is on record as saying SteamOS is going to be an open platform and even the OS its self can be modified... most likely if a developer wants to use Mantle, they can and will.

More than likely they will wind up doing so as they'll already be using it for XBox One, PS4 and PC. The real question is will nvidia ever be able to convince developers to expend time, money and resources coding for a to-the-metal render path for nvidia cards on SteamOS, given that it will only be viable there and will not pay off on every platform, particularly the dominant gaming platform like using Mantile will; consoles.

 

[ShintaiDK]

Considering Gabe Newell is on record as saying SteamOS is going to be an open platform and even the OS its self can be modified... most likely if a developer wants to use Mantle, they can and will.

More than likely they will wind up doing so as they'll already be using it for XBox One, PS4 and PC. The real question is will nvidia ever be able to convince developers to expend time, money and resources coding for a to-the-metal render path for nvidia cards on SteamOS, given that it will only be viable there and will not pay off on every platform, particularly the dominant gaming platform like using Mantile will; consoles.

Its funny consoles keep being mentioned. Yet nowhere is it stated that Mantle is used for consoles.

Not to mention there isnt even tools for Mantle out or a spec for companies to design after. And its not coming anytime soon either as stated by AMD.

 

[sontin]

Huh? There is no problem for nVidia to introduce an own low level API. THX to x86 devs using a high level API to develop their multi port games. They can easily ignore Mantle and use only nVidia's own API. No problem.

 

[3DVagabond]

http://vr-zone.com/articles/mantling-alliances-ritchie-corpus-amd-interview/58215.html

If you think about it, Mantle is truly a collaboration and I’ll tell you [DICE’s Johan Andersson] was at the forefront of that at the very beginning. A lot of the feedback on the development of Mantle came from him. We also solicited feedback from a lot of other partners that we haven’t announced yet. At this stage, Battlefield 4 and FrostBite 3 are the closest to deliver something today. I think, as I mentioned before, the goal would be to provide the spec and SDK publicly.

VRZ: Any timeline to Mantle’s SDK being made public?

It could be as early as sometime next year or maybe the year after.

So, DICE only. No specs or SDK until...who knows when. Not exactly something that will be adopted anytime soon. Assuming it ever will be.

How do you get DICE only from that? They say that they solicited other devs and Frostbite 3 and BF4 is the closest to delivering something, not the only one.

No Mantle for SteamOS.

Not unless Steam box also uses GCN, which it's not going to. This is a close to the metal API for GCN.

 

[3DVagabond]

Not to mention there isnt even tools for Mantle out or a spec for companies to design after. And its not coming anytime soon either as stated by AMD.

They said they weren't making it public yet. Not that there aren't any tools or a spec. Obviously there are. What do you suppose DICE are using?

 

[Grooveriding]

Steam Boxes are also completely hardware flexible except for the low-end box that will be for streaming from your desktop. The next two tiers; better and best are both totally flexible in what hardware builders can choose to put in them. So you'll have some using nvidia cards and others using AMD cards.

Nvidia's challenge, unlike AMD where game developers will already be coding towards GCN parts in consoles, is that they are only in desktops. If a game developer is already using Mantle for work on consoles they can take that and provide the Mantle render path for the port of their game on PC. Considering the financial demands of game development they're not going to expend the time and money to do an additional amount of work to code for an nvidia specific render path that only has use on the least significant platform, PC.

With AMD the work will already have been done on console.

This is going to be very interesting to see how it all plays out. First I want to see the AMD Mantle vs AMD DX11 vs Nvidia DX11 Battlefield 4 benchmarks in December. That is going to be the best indicator of just how powerful, or not, Mantle is going to be for AMD.

 

[sontin]

Nvidia's challenge, unlike AMD where game developers will already be coding towards GCN parts in consoles, is that they are only in desktops. If a game developer is already using Mantle for work on consoles they can take that and provide the Mantle render path for the port of their game on PC. Considering the financial demands of game development they're not going to expend the time and money to do an additional amount of work to code for an nvidia specific render path that only has use on the least significant platform, PC.

With AMD the work will already have been done on console.

And yet devs are using nVidia's GPUs to develop these games. :awe:
It's funny that the same people still ignore the fact that there is no Mantle on any next gen console. And that no developer is using a low level API for a multi port game.

 

[3DVagabond]

@Grooveriding

Yes, I had forgotten that Steambox wasn't fixed hardware. We'll have to see how that plays out. I imagine AMD would have to do something to get Mantle to compile on Steambox. It is supposed to be multi platform compatible though.

Agree with the rest of your post 100%.

 

[Grooveriding]

@Grooveriding

Yes, I had forgotten that Steambox wasn't fixed hardware. We'll have to see how that plays out. I imagine AMD would have to do something to get Mantle to compile on Steambox. It is supposed to be multi platform compatible though.

Agree with the rest of your post 100%.

My guess is that since it's low-level AMD already has planned and taken into account Linux. PS4 is also Linux based and devs will be able to use Mantle there as well for low-level control on GCN.

Really I want to see benchmarks. It's all up in the air until we get to see what sort of difference Mantle can make for PC game performance.