Are Valve and AMD about to ruin PC gaming?

[LtMatt]

I have to admit the article is pretty negative but nonetheless it should get a good debate going.

My biggest gron on this whole affair is with the op. Copy, paste, post and run without adding any input of their own to provoke discussion and debate.

I rarely add my opinion in the OP when i post articles. Its about highlighting a bit of the article and making the sure the article link gets a hit from the end user reading it.

I often end up giving an opinion later in the thread if people choose to discuss it.

I have to admit the article is pretty negative but nonetheless it should get a good debate going.

 

[pgi947]

The article's just a big troll.

This.

The author of that article needs to pull his head out of the sand and look at the (pardon the pun) BIG PICTURE.

Throughout he speaks as if its SteamOS OR Windows and makes mention that one won't support the other and so on.

Does he really REALLY think Valve are stupid enough to alienate 95%+ of their user base by forcing SteamOS? Because that's what it sounds like he's getting at.

SteamOS isn't an out and out replacement for Windows as a gaming platform, its a viable, affordable alternative - at least for now it is. Gabe Newell has the contacts and the respect in the industry to make it work, in the end it won't be a case of 'one or the other' it will be a case of 'either or'. Which ever suits the users needs the most.

As for Mantle, this quote sums up all that Cyril Kowaliski knows about the API:

And because of Mantle, buying a machine with a GPU from the wrong vendor could mean missing out on critical optimizations.

He makes it sound like Nvidia don't make drivers? Is this tool serious? And as we know, its also an open API, absolutely no reason for it ever to get as far as a game flat out not working on a rival GPU. If Nvidia have to swallow their pride and implement mantle they will. Intel did it and still do with X86-64, cross licensing it from AMD, everyone's a winner. Its not in AMD's nature to 'lock out' the competition.

TL;DR: Cyril Kowaliski is a tool that needs to do his homework before publishing nonsensical troll bait articles.

 

[Rusty0611]

With Mantle it depends on how it affects NVidia, if the shoe was on the other foot and NVidia had announced Mantle most of those currently praising it would be in a state of uproar.

I agree.

Although with Mantle it is still a good thing of course.

 

[LtMatt]

How does NVAPI affect AMD? Exactly it doesn't as its built for Nvidia gpu's only. Mantle will be the same in this regard, at least initially.

 

[Lethagized]

With Mantle it depends on how it affects NVidia, if the shoe was on the other foot and NVidia had announced Mantle most of those currently praising it would be in a state of uproar.

The reason we would be in uproar is because we know nVidia would lock it down to their hardware only, or they would pay devs to make it run badly on AMD gear.

 

[Orangey]

Best I can tell only 1 person around here (Rroff) has a good idea what NVAPI actually does so I don't think we can compare it to Mantle just yet.

From what I can tell it doesn't go very far and just provides some small functions for slight speedups here and there.

 

[Final8y]

Best I can tell only 1 person around here (Rroff) has a good idea what NVAPI actually does so I don't think we can compare it to Mantle just yet.

From what I can tell it doesn't go very far and just provides some small functions for slight speedups here and there.

But it would not be a negative if it gave a huge speed up would it.

 

[flopper]

The reason we would be in uproar is because we know nVidia would lock it down to their hardware only, or they would pay devs to make it run badly on AMD gear.

yea the big issue with Nvidia and the woodscrews.
if Nvidia had the console deal that would have had happen big time.

 

[mmj_uk]

How does NVAPI affect AMD? Exactly it doesn't as its built for Nvidia gpu's only. Mantle will be the same in this regard, at least initially.

If NVidia cards get much less performance in Mantle games it means that gamers will be tied into AMD cards. If NVidia decides to create their own CUDA based rival to Mantle, we're going to be in a situation where developers take sides with games optimised for one vendor or another. How is that good for gamers?

 

[no_1_dave]

Mantle will not grow to the scale Glide did back in the day.
I think AMD cards will benefit from Mantle tweaks but DX and OpenGL will still be used as the main APIs.

As for Steam OS..?
Support for it, yes. Used as the main platform that devs focus on, no.

Windows and Direct X are the 2 main things all PCs have so they will carry on being the focus of devs.

 

[drunkenmaster]

If NVidia cards get much less performance in Mantle games it means that gamers will be tied into AMD cards. If NVidia decides to create their own CUDA based rival to Mantle, we're going to be in a situation where developers take sides with games optimised for one vendor or another. How is that good for gamers?

Because developers won't do that, it's that simple. Again this is not the 90's, game dev's aren't all new to 3d gaming, aren't working on tiny budgets, neither are the gpu manufacturers.

Today gpu manufacturers spend millions on dev support and game devs themselves spend, well, up to quarter of a billion on a single game(gta 5, though I expect Blizzard will have spent way more than that on Warcraft).

The idea that Dice would create an engine and attempt to licence it out but skimp on an important API for the sake of relatively little work and making their engine less sell-able is simply wrong.

More performance means more performance, more usage of the hardware you are paying money for is MUCH better than paying for double the performance hardware to make up the difference in lacking performance.

What Mantle will most likely do is either adding Nvidia support at a later date via pushing it to a group to control it outside of AMD, or Mantle will force openGL groups hand to actually get openGL working great, and maybe DX as well.

The lock in to windows and DX has been hurting the industry for a decade, SOMETHING has to happen for that to change. Gamers being given the choice of OS to install, having a better API, better value for money from all their hardware and better games is nothing short of brilliant for gamers. If that is from Mantle itself, Mantle forcing the industry to come up with an alternative, or Mantle forcing MS/openGL group to pull their finger out and make something decent... I honestly don't care.

 

[LtMatt]

If NVidia cards get much less performance in Mantle games it means that gamers will be tied into AMD cards. If NVidia decides to create their own CUDA based rival to Mantle, we're going to be in a situation where developers take sides with games optimised for one vendor or another. How is that good for gamers?

Nvidia cards will still use DX11 API or OpenGL. Mantle will have no effect as it will not be usable on any other card apart from GCN.

 

[Final8y]

Nvidia cards will still use DX11 API or OpenGL. Mantle will have no effect as it will not be usable on any other card apart from GCN.

Indeed Mantle will not hinder DX11 or OpenGL or NV cards from doing what they would be doing anyway.

 

[Syphadeus]

Technically speaking it won't affect Nvidia performance at all as its likely it won't be compatible with their gpu's as its designed for GCN.

I think you're missing the point. Mantle is for "coding to the metal" on AMD GCN hardware, not Nvidia. So the upshot is that PC games with a Mantle renderer option, e.g. BF4 will be introducing AMD specific optimisations that Nvidia owners cannot use.

A possible indirect effect of this is that if Mantle is successful in giving obvious performance improvements and gains traction, it will mean that Nvidias products will not be able to compete because people won't want to buy a product that can't render their game as well as the competition.

The threat of this is greater than previous exclusive features like PhysX and TXAA because the renderer if built from the ground up for the GCN architecture, which is featured in both next next gen consoles and AMDs latest GPUs.

 

[Final8y]

I think you're missing the point. Mantle is for "coding to the metal" on AMD GCN hardware, not Nvidia. So the upshot is that PC games with a Mantle renderer option, e.g. BF4 will be introducing AMD specific optimisations that Nvidia owners cannot use.

A possible indirect effect of this is that if Mantle is successful in giving obvious performance improvements and gains traction, it will mean that Nvidias products will not be able to compete because people won't want to buy a product that can't render their game as well as the competition.

The threat of this is greater than previous exclusive features like PhysX and TXAA because the renderer if built from the ground up for the GCN architecture, which is featured in both next next gen consoles and AMDs latest GPUs.

You have a point but DX and OpenGL are compatibility standards and not for the sake of keeping performance of hardware equal, OpenGL runs better on NV cards, no one is saying that's not fair.

You dont see anyone saying that's its not fair that NV have the fastest single GPU, Intel have the fastest CPU or that the PS4 is potentially better than the XBox1 and that all hardware makers should hold back performance to the slowest competitor for fairness.

And the bold part could be seen as what GPU physx is trying to do, what i personally think about physx is not the point but the fact that someone who is a NV customer is blocked from using GPU physx if they have another make of GPU also in the PC doing the rendering.

 

[mmj_uk]

Nvidia cards will still use DX11 API or OpenGL. Mantle will have no effect as it will not be usable on any other card apart from GCN.

Indeed Mantle will not hinder DX11 or OpenGL or NV cards from doing what they would be doing anyway.

but if Mantle gives AMD cards a huge boost compared to NVidia then NVidia will probably feel compelled to create their own NVidia centric version, if history is anything to go by developers will rarely support both and it will be a battle of which vendor can snap up the most 'exclusive titles'.

 

[Final8y]

but if Mantle gives AMD cards a huge boost compared to NVidia then NVidia will probably feel compelled to create their own NVidia centric version, if history is anything to go by developers will rarely support both and it will be a battle of which vendor can snap up the most 'exclusive titles'.

Which will mean in some titles we will get free boosts and in others we will get standard DX or OpenGL which is better than never getting a free boost.

 

[no_1_dave]

but if Mantle gives AMD cards a huge boost compared to NVidia then NVidia will probably feel compelled to create their own NVidia centric version, if history is anything to go by developers will rarely support both and it will be a battle of which vendor can snap up the most 'exclusive titles'.

It won't give a huge boost.
There will be a boost. We are probably talking a couple of fps at the most.

The selling point of Mantle is there is a lot less overhead for each call vs DX.
DX overhead is dealt with mainly by the CPU (overhead being what the CPU has to do to pass the command onto GPU)
No game apart from BF4 maxes any modern CPU, therefore no game before BF4 would greatly benefit from Mantle.
Any performance gain will be down to minimising the overhead the GPU has to deal with.

Modern CPUs can push through the overhead making it largely irrelevant.
This imo means Mantle will mainly be extra tweaks and some reduced GPU overhead, a bit like NVAPI is for Nvidia but likely to a larger extent leading to slightly increased gains.

 

[Kaapstad]

It won't give a huge boost.
There will be a boost. We are probably talking a couple of fps at the most.

The selling point of Mantle is there is a lot less overhead for each call vs DX.
DX overhead is dealt with mainly by the CPU (overhead being what the CPU has to do to pass the command onto GPU)
No game apart from BF4 maxes any modern CPU, therefore no game before BF4 would greatly benefit from Mantle.
Any performance gain will be down to minimising the overhead the GPU has to deal with.

Modern CPUs can push through the overhead making it largely irrelevant.
This imo means Mantle will mainly be extra tweaks and some reduced GPU overhead, a bit like NVAPI is for Nvidia but likely to a larger extent leading to slightly increased gains.

+1

I am getting on a bit but I have seen this sort of thing many times with technology. A new thing will come along and there will be big claims for increased performance in certain areas but when it comes to real world performance the actual increase is very little.

 

[bru]

The reason we would be in uproar is because we know nVidia would lock it down to their hardware only

Er sorry how is that not what AMD has done. As I understand it Mantle is written to only use AMD cards.

Nvidia cards will still use DX11 API or OpenGL. Mantle will have no effect as it will not be usable on any other card apart from GCN.

Personally I do think that if Nvidia had of announced this then the general opinion of it wouldn't be anyway near as well received as it has been.