Support NeoGAF

[Derpcrawler]

Why would they use 'excellent summary' if it contained any errors? Especially if it were something as massive as the origin of the API being completely wrong.

Nvidia fanboys want to believe that Mantle is new Glide and no one will use it. If it's originate from low level API on Xbox One then it will require zero to no effort for developers to use it in PC versions of the game. I am waiting for 20nm ATi card and switching my GTX780 for that.

 

[KKRT00]

Why would they use 'excellent summary' if it contained any errors? Especially if it were something as massive as the origin of the API being completely wrong.

Having one wrong information that in general doesnt matter much, does not mean that article isnt 'excellent summary'.

 

[satam55]

Last I checked, it's a revised version of libGCM that's used in the PS4. That might have changed, but do you have a source that it's a completely new API?

There's plenty of info in this article:

A more crucial issue is that, while the PS4 toolchain is designed to be familiar to those working on PC, the new Sony hardware doesn't use the DirectX API, so Sony has supplied two of their own.

"The graphics APIs are brand new - they don't have any legacy baggage, so they're quite clean, well thought-out and match the hardware really well," says Reflections' expert programmer Simon O'Connor.


How The Crew was ported to PlayStation 4 (And why PC is so important for the next-gen launch.)

 

[LeleSocho]

If mantle manages to became a little relevant (i doubt so) between this and Steam Machine can we safely say gg consoles?

 

[Brad Grenz]

We're in trouble if linking to an article is supposed to constitute a wholesale endorsement of everything it says. And it's certainly not like AMD was choosing between Anandtech and Eurogamer when they made the tweet.

 

[jwhit28]

If mantle manages to became a little relevant (i doubt so) between this and Steam Machine can we safely say gg consoles?

That makes little sense. AMDs whole hook for developers is that the low level work they are putting in for GCN in the PS4 and XB1 can easily be ported to PC. Including Mantle support makes more sense for console ports than PC only games.

 

[LeleSocho]

That makes little sense. AMDs whole hook for developers is that the low level work they are putting in for GCN in the PS4 and XB1 can easily be ported to PC. Including Mantle support makes more sense for console ports than PC only games.

Yes but if everything works perfectly for them thanks to these two factors pc gaming marketshare would increase and the console one will shrink

 

[mavs]

We're in trouble if linking to an article is supposed to constitute a wholesale endorsement of everything it says. And it's certainly not like AMD was choosing between Anandtech and Eurogamer when they made the tweet.

If it isn't a wholesale endorsement, it's pretty shady to put very positive speculation on your PR campaign when you know it is wrong.

 

[Brad Grenz]

Not really. The article is a good overview of what Mantle does from one of the biggest tech sites on the net. The fact that they might be off the mark with their speculation on something that is ultimately of no import doesn't change that.

 

[Serandur]

Regarding Mantle, if it is specifically, in its current form, only for GCN cards and if it is being implemented only for said GCN cards, when AMD release a new GPU architecture, will the performance advantage be limited to GCN cards thus crippling benchmark performance improvements of the new cards relative to the old in the same games (negating or minimizing the performance increase on those select games)?

 

[childplease]

If mantle manages to became a little relevant (i doubt so) between this and Steam Machine can we safely say gg consoles?

You can sorta say that now. Even without mantle there is just so much brute power on pcs that they will easily have the best version of every next gen game. Right from the very start.

Bf4 will be the first test.

 

[stryke]

Regarding Mantle, if it is specifically, in its current form, only for GCN cards and if it is being implemented only for said GCN cards, when AMD release a new GPU architecture, will the performance advantage be limited to GCN cards thus crippling benchmark performance improvements of the new cards relative to the old in the same games (negating or minimizing the performance increase on those select games)?

Kind of a moot point when they'll just create Mantle 2.0 or use DirectX instead.

 

[Serandur]

Kind of a moot point when they'll just create Mantle 2.0 or use DirectX instead.

The problem with Mantle 2.0 being that it's not being coded for right now, and if I understand it correctly, Mantle's current integration is able to be a low-level API specifically because it only targets a specific GPU architecture, therefore current iterations may not benefit future architectures (?). I'm not asking whether future GPUs won't have a version of Mantle or won't be able to run DirectX games, I'm just asking if, say for example, AMD's next line of GPUs are not GCN-based (just hypothetical) and have a general typical performance improvement of about 30% if that performance improvement will be minimized with certain games based on incompatibility with the version of Mantle utilized? I'm not sure if that makes any sense, but the limited understanding I have of Mantle is that it is a very specifically designed API that simply will not work, as it currently is, with future GPU architectures and therefore the older cards have a performance boost with games using it relative to future cards. Is this incorrect?

 

[gofreak]

The problem with Mantle 2.0 being that it's not being coded for right now, and if I understand it correctly, Mantle's current integration is able to be a low-level API specifically because it only targets a specific GPU architecture, therefore current iterations may not benefit future architectures (?). I'm not asking whether future GPUs won't have a version of Mantle or won't be able to run DirectX games, I'm just asking if, say for example, AMD's next line of GPUs are not GCN-based (just hypothetical) and have a general typical performance improvement of about 30% if that performance improvement will be minimized with certain games based on incompatibility with the version of Mantle utilized? I'm not sure if that makes any sense, but the limited understanding I have of Mantle is that it is a very specifically designed API that simply will not work, as it currently is, with future GPU architectures and therefore the older cards have a performance boost with games using it relative to future cards. Is this incorrect?

I don't see the big problem in releasing a new API for each new generation of architecture, if they want to do that. Every couple of years or whatever.

Engines would 'just' add another render path for the new API while keeping the render path for Mantle 1.0 around for older hardware.

 

[Serandur]

I don't see the big problem in releasing a new API for each new generation of architecture, if they want to do that. Every couple of years or whatever.

Engines would 'just' add another render path for the new API while keeping the render path for Mantle 1.0 around for older hardware.

No big problem, but would those older games and Mantle 1.0 also benefit the newer cards? That's what I'm asking. Again, I don't think it's to be a big issue, just wondering.

 

[AdmiralPain]

Anand is very reliable, the crew there know their shit. I believe them.

 

[Brad Grenz]

They admit it's a wild guess, but whatever.

 

[SparkTR]

I really hope Nvidia start supporting this, and AMD starts a Linux push in the not too distant future. There seems to be a lot of interesting stuff going on in the PC space, now we just need some unity.

 

[gofreak]

No big problem, but would those older games and Mantle 1.0 also benefit the newer cards? That's what I'm asking. Again, I don't think it's to be a big issue, just wondering.

Good q...probably not as much, no. Without patches anyway.

But might be a moot point since older games will increasingly blaze along on newer and newer hardware via regular DX or OGL anyway. Performance utilisation / 'efficiency' might be more important on contemporary rather than legacy games.

 

[TheLegendary]

Any chance this is what cboat or someone was referencing that MS wasn't able to announce anything until AMD cleared it later on (theoretically end of September)?

 

[Datschge]

To me it looks like this right now: AMD stated that they were working on Mantle for PC for 2 years. Sony on PS4 essentially announced a successor to libGCM on their own, focussing PR on the ability to code to the metal. Microsoft previously stated that everything on the XBone would go through an adapted DirectX layer, which backfired at that point. Since AMD already was working on a lower level API on the PC and Microsoft needed a "code to the metal" solution fast to counter Sony's one they ported Mantle.

But in the end it may not even matter all that much, in all three cases they are low level APIs interfacing the same hardware backend. Even if their API development was independent, AMD's (as the hardware designer) input likely ensured that they are comparable enough to be compatible with few tweaks.

 

[kinggroin]

No big problem, but would those older games and Mantle 1.0 also benefit the newer cards? That's what I'm asking. Again, I don't think it's to be a big issue, just wondering.

Would you care honestly? If you're playing older Mantle 1.0 games on new hardware, you're probably running those games just as good if not better than you did back on release.

Like you said, its not a big issue.

 

[Serandur]

Would you care honestly? If you're playing older Mantle 1.0 games on new hardware, you're probably running those games just as good if not better than you did back on release.

Like you said, its not a big issue.

simply curiosity for curiosity's sake

 

[gofreak]

To me it looks like this right now: AMD stated that they were working on Mantle for PC for 2 years. Sony on PS4 essentially announced a successor to libGCM on their own, focussing PR on the ability to code to the metal. Microsoft previously stated that everything on the XBone would go through an adapted DirectX layer, which backfired at that point. Since AMD already was working on a lower level API on the PC and Microsoft needed a "code to the metal" solution fast to counter Sony's one they ported Mantle.

But in the end it may not even matter all that much, in all three cases they are low level APIs interfacing the same hardware backend. Even if their API development was independent, AMD's (as the hardware designer) input likely ensured that they are comparable enough to be compatible with few tweaks.

Microsoft licensing Mantle for a low level Xbox One API certainly would be more plausible than the opposite.

However in that case, the two APIs would still diverge. Microsoft would need specific things for their API not in Mantle (relating to XB1's memory setup).

Microsoft also characterised the Mono driver as a stripped down DX variant though, rather than a GCN from-scratch API.

 

[Captain Tuttle]

If AMD thought the article was way off base I don't think they'd tweet that it was an "excellent summary". If some of the speculation was wrong I think they'd just ignore the story like companies usually do.

 

[kartu]

Yes but if everything works perfectly for them thanks to these two factors pc gaming marketshare would increase and the console one will shrink

Uh, why?
If you mean SteamBox, first I so doubt it will take off and second, why would AMD care?

One thing this would do for sure is, gamers will tend to choose AMD's GPUs,.

 

[Datschge]

However in that case, the two APIs would still diverge. Microsoft would need specific things for their API not in Mantle (relating to XB1's memory setup).

Wouldn't memory handling be outside of Mantle's API though? Even on PC Mantle would eventually need to work with different memory structures in GCN being on discrete cards and as part of APUs so I'd expect that part to be handled more abstracted on a higher level and be an implementation detail in the underlying driver. Or something like that.

Microsoft also characterised the Mono driver as a stripped down DX variant though, rather than a GCN from-scratch API.

Maybe they are just referring to it being compatible with DX's HLSL? Then using the above wording is a great way to save face. =D

 

[gofreak]

Wouldn't memory handling be outside of Mantle's API though? Even on PC Mantle would eventually need to work with different memory structures in GCN being on discrete cards and as part of APUs so I'd expect that part to be handled more abstracted on a higher level and be an implementation detail in the underlying driver. Or something like that.

Depends on the scope of the API, but if we're talking about something low level, I think there'd need to be exposure around what memory is being used for certain texture/buffer related operations in XB1, things that are taken for granted in a PC memory model.

Maybe they are just referring to it being compatible with DX's HLSL? Then using the above wording is a great way to save face. =D

When they talked about it, they said it was a case of starting with DX and taking things out and optimising the remainder for the chip.

That doesn't sound like shader syntax compatibility.

Like I said, I could see AMD licensing Mantle to MS, but couldn't see MS licensing something like the above for open cross platform use, under AMD's control. Of course it's possible MS isn't telling the whole truth about the genesis of their driver, and that they did do a license of Mantle, but it would be a modified version even in that case methinks rather than something the exact same as the PC version.


...

It's fairly academic though. It doesn't really matter, except that it would make PC versions of console games even easier to optimise for AMD hardware, if there was a direct analog in the console space. But I don't think Anandtech supports their speculation well, there's lots of other explanations for HLSL syntax compatibility.

 

[dr_rus]

Nvidia fanboys want to believe that Mantle is new Glide and no one will use it. If it's originate from low level API on Xbox One then it will require zero to no effort for developers to use it in PC versions of the game. I am waiting for 20nm ATi card and switching my GTX780 for that.

You're ready to play straight console ports on your PC then? That's as far as you go with zero effort.

 

[Durante]

Nvidia fanboys want to believe that Mantle is new Glide and no one will use it.

I'm old enough to remember running Glide games, and it certainly was used quite a lot at the height of 3Dfx' power. It also worked really well in those games that supported it, only when you were using 3DFx hardware of course.

However, you seem to be saying that Mantle is different. But really, how is Mantle not like Glide? It's a low-level vendor-exclusive 3D API for PC. Its concept seems to match Glide to a T.

 

[SapientWolf]

I'm old enough to remember running Glide games, and it certainly was used quite a lot at the height of 3Dfx' power. It also worked really well in those games that supported it, only when you were using 3DFx hardware of course.

However, you seem to be saying that Mantle is different. But really, how is Mantle not like Glide? It's a low-level vendor-exclusive 3D API for PC. Its concept seems to match Glide to a T.

Glide wasn't used on any major consoles. That's the main difference, I guess.

 

[KidBeta]

Glide wasn't used on any major consoles. That's the main difference, I guess.

And theres no confirmation that mantle is either.

 

[sirap]

Yeahh, just because the new consoles use AMD hardware doesn't mean Nvidia's going to lay down and do nothing in the PC space. You can bet they'll fight tooth and nail against this or come out with their own alternative.

 

[SapientWolf]

And theres no confirmation that mantle is either.

OP jumped the gun? If it's true I'm surprised the info didn't leak earlier. If it's not true then I'm surprised that DICE is willing to do all that low level API work on the PC side if they can't reuse any code from the console ports.

 

[Laughing_Man__]

*Raises hand*

QUESTION!

Why would AMD link that if it was NOT true about the X1?

 

[Lazaro]

http://pc.watch.impress.co.jp/img/pcw/docs/593/760/15.jpg

Isn't it possible that this could be Microsoft/AMDs idea of expanding an evolution of DirectX that runs on multiple systems. You know to undermine OpenGL?

Sounds stupid but it's possible.

 

[DeFiBkIlLeR]

Mantle has me worried on the hardware side of things, if they can get an extra 20,30,40% extra performance out of current GPU's (or whatever the figure may be) just by using improved driver software...where's the incentive to bring out genuinely new cards using smaller fabrications and lower power usage?

What they used to have to do with new generations of cards every 12-18mths or so, can now be achieved in software at a relatively tiny cost.

GPU stagnation could be on the way....

 

[Gamesadict]

*Raises hand*

QUESTION!

Why would AMD link that if it was NOT true about the X1?

Thing is people are immediately assuming the X1 part to be 100% accurate, when that is just one part of a bigger article. Besides, the tweet mentions it as a good explanation of the architecture, there is no mention of its origin. I pretty much re-worded what someone already said.

 

[SapientWolf]

Mantle has me worried on the hardware side of things, if they can get an extra 20,30,40% extra performance out of current GPU's (or whatever the figure may be) just by using improved driver software...where's the incentive to bring out genuinely new cards using smaller fabrications and lower power usage?

What they used to have to do with new generations of cards every 12-18mths or so, can now be achieved in software at a relatively tiny cost.

GPU stagnation could be on the way....

I have some bad news for you...

 

[aeolist]

Mantle has me worried on the hardware side of things, if they can get an extra 20,30,40% extra performance out of current GPU's (or whatever the figure may be) just by using improved driver software...where's the incentive to bring out genuinely new cards using smaller fabrications and lower power usage?

What they used to have to do with new generations of cards every 12-18mths or so, can now be achieved in software at a relatively tiny cost.

GPU stagnation could be on the way....

gpu stagnation is already here thanks to slowing process node shrinks and hitting the practical limits of die size and power draw

but even when there's no real change every year amd and nvidia have to launch a "new" series to please oems that want bigger model numbers in their pcs

there will still be jumps when tsmc/glofo get to 20nm and beyond but i imagine the architectures are pretty well-optimized and won't be seeing any huge redesigns in the next few years. competition is still intense though and both big players will be trying to get marketshare however they can.

 

[kartu]

GPU stagnation could be on the way....

Sorry, couldn't help:

 

[vg260]

This is really confusing taken in with all the SteamOS stuff. Does all this imply we could be seeing different platforms on the PC even within the Windows-based stuff. I don't really follow the Steam OS stuff either, and this may be off-topic, but could there be exclusives in the PC space moving forward? (I guess that's already the case though if you don't run Windows).

 

[aeolist]

This is really confusing taken in with all the SteamOS stuff. Does all this imply we could be seeing different platforms on the PC even within the Windows-based stuff. I don't really follow the Steam OS stuff either, and this may be off-topic, but could there be exclusives in the PC space moving forward? (I guess that's already the case though if you don't run Windows).

you mean gpu exclusives?

everyone will still make all of their pc games directx and/or opengl compatible for maximum sales. if it's got an opengl rendering path adding os x/linux/steam os support is probable.

mantle (for now) is just a nice bonus for amd owners on certain windows games, though it could easily be added to amd's linux drivers (assuming they decide to start putting effort into their linux drivers in the first place). apple's lockdown on os x and their extremely slow update cadence makes mantle support on that platform unlikely.

 

[moniker]

I know nothing about this stuff, so maybe this is a dumb question. Will devs be able to mix high level and low level code (similar to sprinkling your C code with a bit of ASM where there are bottlenecks) - or would the whole renderer need to be either D3D/OGL or Mantle?

I'm guessing the latter.

 

[Horse Armour]

This is really confusing taken in with all the SteamOS stuff. Does all this imply we could be seeing different platforms on the PC even within the Windows-based stuff. I don't really follow the Steam OS stuff either, and this may be off-topic, but could there be exclusives in the PC space moving forward? (I guess that's already the case though if you don't run Windows).

AMD are the smallest graphics vendor so there's no chance anyone creates a Mantle exclusive game.

 

[dr_rus]

Glide wasn't used on any major consoles. That's the main difference, I guess.

Glide allowed exactly the same at its time - easy portability from N64 and PS1 to PC. This was one of the reason why it was successful at first.

 

[slapnuts]

That's unlikely going to happen, I know a ton of buddies with consoles that would gladly have a box designed for gaming by steam because of the entire back catalog of when they used to PC game. Also a buddy of mine that has an older gaming PC doesn't have the time and money to build a new power machine so having steam do it for him at an affordable price is killer. Steam Machine is going to be a huge step forward this generation.

I don't think so imho...

 

[ShadyJ]

all dem APIs...