The AMD Mantle Thread

[Saylick]

DX10-11 would essentially have to be the biggest turd alive for that to be true. Not to mention how DX9 would be. Or how big an epic flop OpenGL would have to be. All fixed by magic fairies to gain minimum 30-50% on a new API.

There is a reason why AMD didnt show any real gain numbers. Its nothing great to write home about.

Will it be better than DX? Most likely. But its not gonna be near those hype numbers.

AMD's silence = Mantle is a flop VS. Carmack and other developer's word that DX overhead sucks.

One of these is opinion. The other is concrete.

I agree that marketing in general should be taken with a huge grain of salt, but AMD could be withholding numbers because maybe Mantle is nothing to be hyped about, or they could be withholding numbers because they want to save it for November.

If Mantle is indeed the XBO's API, then it's nothing more than a hyped up API port over to the PC. I surmise it's been in development for the past couple of years only because BF4 was in development for next-gen consoles, so it's natural that there wasn't much work performed to get the XBO API in PC-usable form. Using a low overhead API would benefit consoles more than PCs since consoles have a lot less hardware and I would agree that Mantle is worthless ONLY if the DX overhead wasn't that bad. However, more than one person has criticized the vast gap between theoretical performance and actual performance as a result of using a high-level API, one of which is AT.

So whose word do you believe?

 

[Gikaseixas]

Now, why exactly are you so certain it won't be any good?

Giving that this comes from the same source as - Xbone and PS4 won't have an APU

If it works as intended it will be a major boost IMO and judging by the enthusiasm aroujnd Mantle it might be the real deal. Let's hope so.
Another thing that bothers me a bit, Nvidia cannot stay out on this one since it is a major player in the graphics area. They will surely come up with their own version if it succeeds.

 

[lamedude]

There's a reason Direct3D/OpenGL became the standard and hardware specific APIs died. People cheering this announcement solely based on claims of increased performance are being short sighted.

I look forward to beta testing Mantle wrappers so I can use those Mantle exclusive features on my Geforce 1880 and RadeonR9 990. I'm going stock up on torches and pitchforks. As bad as the complaints against Nvidia for "paying" devs to use NVAPI to deprive AMD users of something as simple as AA were I'm expecting a lynch mob when AMD returns the favor in spades.

 

[ShintaiDK]

AMD's silence = Mantle is a flop VS. Carmack and other developer's word that DX overhead sucks.

One of these is opinion. The other is concrete.

I agree that marketing in general should be taken with a huge grain of salt, but AMD could be withholding numbers because maybe Mantle is nothing to be hyped about, or they could be withholding numbers because they want to save it for November.

If Mantle is indeed the XBO's API, then it's nothing more than a hyped up API port over to the PC. I surmise it's been in development for the past couple of years only because BF4 was in development for next-gen consoles, so it's natural that there wasn't much work performed to get the XBO API in PC-usable form. Using a low overhead API would benefit consoles more than PCs since consoles have a lot less hardware and I would agree that Mantle is worthless ONLY if the DX overhead wasn't that bad. However, more than one person has criticized the vast gap between theoretical performance and actual performance as a result of using a high-level API, one of which is AT.

So whose word do you believe?

I think it speaks for itself. When we see the Xbox One already not having enough GPu and CPU power. And soem games being released at 900p30 with dynamic downscaling to go further down. Simply because there isnt enough power to drive it. So if Mantle is the Xbox One API, then its really nothing to write home about and wont deliver those 30%-250% boost as some is already far away in dreamland about.

 

[SiliconWars]

How do you even come up with those numbers?

Do you really think there is even 30%-50% to be gained? And not more in the area of say 5% or less?

This is a classic hype.

Oh please, 5% or less? Do you really believe AMD would spend years developing something that only gives them 5% increase in games that adopt it? And what about repi's _very_ high performance tweet?

You're dreaming if you think 5% is the likely result, I'm putting 20% at an absolute minimum.

 

[Saylick]

I think it speaks for itself. When we see the Xbox One already not having enough GPu and CPU power. And soem games being released at 900p30 with dynamic downscaling to go further down. Simply because there isnt enough power to drive it. So if Mantle is the Xbox One API, then its really nothing to write home about and wont deliver those 30%-250% boost as some is already far away in dreamland about.

Fair point. If consoles are currently using some form of Mantle, i.e. a low-overhead API, and are still struggling to meet performance targets, wouldn't the situation be worse without Mantle? If a properly utilized low overhead API at the console level isn't enough to drive good framerates, then the issue is either A) there simply isn't enough hardware, B) the level of detail developers want to achieve is just too high for the given hardware, or C) a combination of both. In other words, wouldn't developers try to squeeze out every ounce of power that they can out of these systems? If the software optimizations can't get them to their performance targets, then the fault lies in the hardware. Either way, the software optimizations, i.e. Mantle, is beneficial and if these optimizations can be transferred to a PC setting, we should see PC performance go up as well.

I guess the question comes down to how much does it help on consoles, and more importantly, how much of that performance increase is transferable to PCs.

 

[raghu78]

Oh please, 5% or less? Do you really believe AMD would spend years developing something that only gives them 5% increase in games that adopt it?

well said.

And what about repi's _very_ high performance tweet?

You're dreaming if you think 5% is the likely result, I'm putting 20% at an absolute minimum.

yeah. 20 - 25% minimum and more likely 40 - 50%. Given the feedback from developers like John Carmack that console hardware can match upto 2x as fast PC hardware I think its easily possible .

 

[StrangerGuy]

Oh please, 5% or less? Do you really believe AMD would spend years developing something that only gives them 5% increase in games that adopt it? And what about repi's _very_ high performance tweet?

You're dreaming if you think 5% is the likely result, I'm putting 20% at an absolute minimum.

5% for AMD is still better than 0% for Intel HD whatever.

I wonder why somebody here is so indignant about the whole thing. So what are the chances of Intel supporting Mantle? Zero?

 

[Arkadrel]

Do you really think there is even 30%-50% to be gained? And not more in the area of say 5% or less?

I know people are allowed to speculate on stuff, and that you dont like AMD.

However the 5% stuff sounds like a brainfart number.
Think about it ShintaiDK.

If it was that low, you think it would even be worth the effort?
instead of not just staying with Directx?

Codeing to metal gives more than that, its the reason consols are so powerfull with small hardware vs pcs.

 

[sontin]

5% for AMD is still better than 0% for Intel HD whatever.

I wonder why somebody here is so indignant about the whole thing. So what are the chances of Intel supporting Mantle? Zero?

Mantle is GCN only.
I mean how many companies supported Glide? :\

 

[Maxima1]

Oh please, 5% or less? Do you really believe AMD would spend years developing something that only gives them 5% increase in games that adopt it?

Bulldozer.

Though, I agree that 5% doesn't make sense.

 

[StrangerGuy]

Mantle is GCN only.
I mean how many companies supported Glide? :\

Yeah, how many PS1/2s, N64s and Segas had GCN either?

 

[sontin]

3DFx lost the dreamcast deal to PowerVr...

 

[Erazor51]

It will be way more then 5%, Raja wouldn't have said "it's like getting 2018 hardware today" if there wasn't a big performance gain.

 

[sandorski]

If it is a superior API, offers significant performance enhancements, and is easy to port to PC, it won't matter to Developers whether AMD has a large PC Marketshare. They'll simply include it as part of the port. That will likely push AMD PC sales, GPUs for certain, perhaps even CPU sales.

It will all depend on what it is, how capable it is, and how well it works in comparison to other APIs.

 

[SiliconWars]

Bulldozer.

Though, I agree that 5% doesn't make sense.

Lol true about BD - difference here is they had to release that, they don't need to release Mantle.

The way I see it, AMD wouldn't even consider putting this amount of effort in to anything that didn't give them 15-20% at a minimum. Really we're looking at 50%+ I feel, or maybe even more. The language used by everyone involved in it says it all.

We've had Titan "ridiculed" in BF4, "2018 performance today" and "_very_ high performance" from a guy who doesn't even work for AMD, but knows exactly how the performance is better than anyone else.

Yup, 5% is just not happening and I doubt my very low-ball 20% is all we'll get either.

 

[flopper]

Lol true about BD - difference here is they had to release that, they don't need to release Mantle.

The way I see it, AMD wouldn't even consider putting this amount of effort in to anything that didn't give them 15-20% at a minimum. Really we're looking at 50%+ I feel, or maybe even more. The language used by everyone involved in it says it all.

We've had Titan "ridiculed" in BF4, "2018 performance today" and "_very_ high performance" from a guy who doesn't even work for AMD, but knows exactly how the performance is better than anyone else.

Yup, 5% is just not happening and I doubt my very low-ball 20% is all we'll get either.

Mantle seems to be a gamechanger to allow amd to support cpu/gpu with the idea you dont need Intel cpu anymore to game at X fps to maximize the gaming experience.
if you can set up a gaming rig at half the price with as good experience, its a no brainer.

 

[Silverforce11]

That will likely push AMD PC sales, GPUs for certain, perhaps even CPU sales.

Not CPU, it will help AMD's APU tremendously but their CPU will still be crap compared to Intel.

An Intel CPU + CGN is still going to be better than AMD CPU + GCN even with Mantle.

Think of coding to the metal as a hardware power multiplier. Crap multiplied a few times is still gonna be crap in comparison to awesome multiplied a few times.

 

[blackened23]

AMD's silence = Mantle is a flop VS. Carmack and other developer's word that DX overhead sucks.

One of these is opinion. The other is concrete.

I agree that marketing in general should be taken with a huge grain of salt, but AMD could be withholding numbers because maybe Mantle is nothing to be hyped about, or they could be withholding numbers because they want to save it for November.

If Mantle is indeed the XBO's API, then it's nothing more than a hyped up API port over to the PC. I surmise it's been in development for the past couple of years only because BF4 was in development for next-gen consoles, so it's natural that there wasn't much work performed to get the XBO API in PC-usable form. Using a low overhead API would benefit consoles more than PCs since consoles have a lot less hardware and I would agree that Mantle is worthless ONLY if the DX overhead wasn't that bad. However, more than one person has criticized the vast gap between theoretical performance and actual performance as a result of using a high-level API, one of which is AT.

So whose word do you believe?

You will never get an objective answer to the last question. It's a common theme from folks here that have an extreme bias for one brand.

Industry heavy hitters such as John Carmack, Timothy lottes, among others have openly lamented the API situation on windows being terrible while leaving a ton of performance on the table. So the real question is, is shintaiDK a better developer than they are?
This is what Timothy Lottes stated:

The real reason to get excited about a PS4 is what Sony as a company does with the OS and system libraries as a platform, and what this enables 1st party studios to do, when they make PS4-only games. If PS4 has a real-time OS, with a libGCM style low level access to the GPU, then the PS4 1st party games will be years ahead of the PC simply because it opens up what is possible on the GPU. Note this won't happen right away on launch, but once developers tool up for the platform, this will be the case. As a PC guy who knows hardware to the metal, I spend most of my days in frustration knowing damn well what I could do with the hardware, but what I cannot do because Microsoft and IHVs wont provide low-level GPU access in PC APIs. One simple example, drawcalls on PC have easily 10x to 100x the overhead of a console with a libGCM style API.

I for one do not believe him, even though Mr. Lottes worked for nvidia. After all, the former lead software engineer for nvidia (Tim Lottes) who now is a developer @ Epic Games is clearly a moron.

 

[SiliconWars]

Not CPU, it will help AMD's APU tremendously but their CPU will still be crap compared to Intel.

An Intel CPU + CGN is still going to be better than AMD CPU + GCN even with Mantle.

Think of coding to the metal as a hardware power multiplier. Crap multiplied a few times is still gonna be crap in comparison to awesome multiplied a few times.

That's not really what should happen. One part of mantle allows for 8 cores to work if required, and reducing the draw call CPU penalty should remove the AMD CPU bottleneck.

If it works as planned, CPU's will become near irrelevant to overall performance in many games as the bottleneck shifts to the GPU in almost all cases.

 

[Zstream]

And Intel/nVIDIA combined holds the majority of the PC GPU market. I have no issues with consoles, but its the idea that this API will try to compete against DX and OGL as a replacement is the issue I see because more than half the user base could may well be alienated.

I read somewhere that mantle is an open source standard (need citation), but im thinking its openness is more akin to CUDA being "open"..

Quit shifting arguments. Just face it, this was a risk Intel and Nvidia took by not lowering margins for the console. To bad, to sad, get over it, they're big boys.

 

[Zstream]

If it is a superior API, offers significant performance enhancements, and is easy to port to PC, it won't matter to Developers whether AMD has a large PC Marketshare. They'll simply include it as part of the port. That will likely push AMD PC sales, GPUs for certain, perhaps even CPU sales.

It will all depend on what it is, how capable it is, and how well it works in comparison to other APIs.

Hope that wasn't sarcasm but that is very well stated.

 

[frozentundra123456]

Not CPU, it will help AMD's APU tremendously but their CPU will still be crap compared to Intel.

An Intel CPU + CGN is still going to be better than AMD CPU + GCN even with Mantle.

Think of coding to the metal as a hardware power multiplier. Crap multiplied a few times is still gonna be crap in comparison to awesome multiplied a few times.

Not so sure it will help apus as much. The point of mantle seems to be to lower CPU requirements. Apus are bandwidth limited, so I don't know how it would affect that. Maybe it would lower bandwidth requirements as well.

 

[cmdrdredd]

Quit shifting arguments. Just face it, this was a risk Intel and Nvidia took by not lowering margins for the console. To bad, to sad, get over it, they're big boys.

It is also why both are doing better than amd.

 

[cmdrdredd]

Fair point. If consoles are currently using some form of Mantle, i.e. a low-overhead API, and are still struggling to meet performance targets, wouldn't the situation be worse without Mantle? If a properly utilized low overhead API at the console level isn't enough to drive good framerates, then the issue is either A) there simply isn't enough hardware, B) the level of detail developers want to achieve is just too high for the given hardware, or C) a combination of both. In other words, wouldn't developers try to squeeze out every ounce of power that they can out of these systems? If the software optimizations can't get them to their performance targets, then the fault lies in the hardware. Either way, the software optimizations, i.e. Mantle, is beneficial and if these optimizations can be transferred to a PC setting, we should see PC performance go up as well.

I guess the question comes down to how much does it help on consoles, and more importantly, how much of that performance increase is transferable to PCs.

Consoles don't use mantle at all. People keep talking like they do but they don't. Games that are in development are not using it. They have their own api currently in use that does the sane thing but can't be ported. Seriously Microsoft said that dx11 on the Xbox one is built to use every drop of the system and Sony has their own api that does the same thing.