NVIDIA’s GameWorks Blocking AMD From Optimizing?

I could be wrong, but I've always been under the assumption that nvidia works more closely with game developers.

Apparently NVIDIA worked a lot closer with Valve to make SteamOS happen too. Now I'm sure one day benchmarks will show that NVIDIA hardware performs better on SteamOS than comparable AMD hardware... and no doubt people will complain that this is all because of some unfair advantage that NVIDIA coerced out of Valve. When the reality is that NVIDIA decided to show up while AMD was twiddling their thumbs. You know, the same reason why Intel stomps AMD in the CPU space.
 
I thought I read an article wherein AMD could make use of the G-Sync chip by coding it into their own drivers. Or am I mistaken
 
Continuing to say something that is wrong doesn't make it right.

“The plan is, long term, once we have developed Mantle into a state where it’s stable and in a state where it can be shared openly [we will make it available]."

It's not at all the same situation as GameWorks, G-Sync, PhysX, etc.

Because Mantle is not any situation at all -- its vapor at this point. And vague, pie in the sky assertions about Mantle "being shared openly" (with competitors) when the day comes that its "stable" is laughable.

These companies don't support competitors initiatives because that's just the arrogant corporate way. Intel and NVIDIA would never support an API that AMD has sole control of. And vice versa.
 
I never said AMD doesnt do some shitty stuff. Also I would like to see AMD on top for once as nvidia has been dominating the market for far too long dictating stupid high pricing whenever they want. I know that at least for the better part of a decade nvidia has had more market share and sales than ATi/AMD in the graphics sector. I would like AMD for once to be either 50% or a little higher so nvidia can't do $800 GPUs anymore. I am also curious to see what AMD will do when they are finally on top.

I would like for once that I find a lucky pot of gold at the end of every rainbow, and I would like for once that AMD writes a decent driver that isn't still in beta months after major AAA titles launch even when they're partnered for that game.

The market dictates who has the most market share and can charge more. NVIDIA does not dictate "stupid high pricing" - the market does, in part due to the lack of solid alternative competition.
 
Because Mantle is not any situation at all -- its vapor at this point. And vague, pie in the sky assertions about Mantle "being shared openly" (with competitors) when the day comes that its "stable" is laughable.

These companies don't support competitors initiatives because that's just the arrogant corporate way. Intel and NVIDIA would never support an API that AMD has sole control of. And vice versa.

That's funny because the green team is throwing Mantle around like its the biggest threat to pc gaming. So what is it, vaporware or real? When it comes out are you going to be laughing then or crying in your green tea?
 
Ever heard of the AMD-64 instruction set? Ever heard of the Havok Physics engine?

Yes, because those two examples are obviously representative of every R&D effort of the entire technology industry.

Except that they aren't.

Until we get some third party organization that manages to get both Nvidia and AMD on board to develop a set of universal standards, both companies are going to keep working on proprietary code in one form or another.

If 3dfx were still around, and Matrox were doing something relevant to gaming, maybe. But that's just not the case, and I seriously wouldn't expect anyone to throw millions of dollars into a volcano to try and compete with AMD and Nvidia on every level from $200-$500 cards right now. Instead we're basically left with 2 juggernauts that do not want to work together, no matter how much one or the other might try to portray themselves as wanting to be the cooperative one in the relationship at any given time.
 
Until we get some third party organization that manages to get both Nvidia and AMD on board to develop a set of universal standards, both companies are going to keep working on proprietary code in one form or another.

We already have that. It's called the Khronos Group. But developers are under no obligation to use open / industry standards.
 
Mantle is an API for AMD's GCN architecture.



And it is proprietary.

Mantle expects a certain hardware feature set as a baseline (which they're obviously basing on GCN chips). Mantle implements an extension system for this reason, vendor specific shit is meant to be implemented as such.
 
Hey guys AMD needs all the help it can get when optimizing it's drivers.

I mean what else is mantel but a cry for help from AMD for developers to pre-optimize games for AMD cards. Mantel is open because it doesn't matter, it's designed around AMD arch GCN and any changes to mantel will be to benefit AMD first. Yeah sure nvidia is at liberty to use mantel but it's only going to really benefit AMD first and foremost.
 
We already have that. It's called the Khronos Group. But developers are under no obligation to use open / industry standards.

I meant a third party group that's relevant and anyone gives a damn about. Khronos Group? Never heard of it. Their website isn't hard to find, sure there are AMD and Nvidia right there as promoter members. Let's see, is this group all about setting up one standard? Nope, there's 15 listed at the top of the page along with another section for "other". Looking around some more... ok, so it seems the standards are for various things and there's a few that are related to 3d rendering... Oh, look at that, OpenGL! That thing no one has really cared about for years. That group us pretty useless.

We need a new group. Not one run by 2 companies(Nvidia and AMD), not one that seems to have a meandering sense of direction since it wants to do everything from opengl to media portability. We need ONE that is focused on an API, that's not run by Microsoft, and isn't simply trying to keep OpenGL from falling into obscurity. About the only other company that might possibly throw wads of cash into the GPU development volcano sometime in the next decade is Intel(and it doesn't count if they just buy Nvidia), so again with no serious competition, the two juggernauts in the marketplace(AMD and Nvidia) are not going to cooperate and actually produce something that benefits everyone.
 
Hey guys AMD needs all the help it can get when optimizing it's drivers.

I mean what else is mantel but a cry for help from AMD for developers to pre-optimize games for AMD cards. Mantel is open because it doesn't matter, it's designed around AMD arch GCN and any changes to mantel will be to benefit AMD first. Yeah sure nvidia is at liberty to use mantel but it's only going to really benefit AMD first and foremost.

Before trying to sound like you know something about anything, please try to get the spelling right. It's Mantle.

Now go away and actually read (yes I know, hard stuff) about what Mantle is, then come back and comment.

Mantle is, in fact a response to a cry for help from developers that are sick of having to deal with the crud API that MS has foisted on PC gaming for over a decade. It will be available to other vendors, AMD has a history of trying for open standards - if you run an R9 290 and a 780 in the same machine, Mantle will work on the R9 regardless of waht nVidia do.

Can you say the same about PhysX?
 
From what I've seen over the past few years, and what I've heard from developers who choose to come out and say it, is that AMD's engineering help pales in comparison to NVIDIA's. Sometimes they are lucky to even get a call returned from AMD while NVIDIA sends an engineer out the next day cross-country to the developers studio, etc.

Sounds like the game studios just want NVIDIA's quality of support yet also want NVIDIA to optimize their game for NVIDIA's competitor as well. Screw that.
 
I thought I read an article wherein AMD could make use of the G-Sync chip by coding it into their own drivers. Or am I mistaken

Not exactly, but an AMD exec did say they've been working on their own solution along with 4k integration and new VESA standards.
 
Before trying to sound like you know something about anything, please try to get the spelling right. It's Mantle.

Now go away and actually read (yes I know, hard stuff) about what Mantle is, then come back and comment.

Mantle is, in fact a response to a cry for help from developers that are sick of having to deal with the crud API that MS has foisted on PC gaming for over a decade. It will be available to other vendors, AMD has a history of trying for open standards - if you run an R9 290 and a 780 in the same machine, Mantle will work on the R9 regardless of waht nVidia do.

Can you say the same about PhysX?
cool so you can read press packets good job yeah everyone knows those are reliable. it's a low level api designed to bring out the best of GCN do i need to say more? It's a pseudo open standard, it's like saying google's android is a open platform... corporate open and community driven open are two very different things.

No fucking shit mantle will work on a AMD card regardless of what nvidia does; what kind of statement is that. Are you suggesting that nvidia broke openGL/DirectX in a game making it impossible to run on a AMD card? Else that statement makes no sense. Can i say the same about physX yes it's easy " PhysX will work on a 780 regardless of what AMD does."
 
No fucking shit mantle will work on a AMD card regardless of what nvidia does; what kind of statement is that. Are you suggesting that nvidia broke openGL/DirectX in a game making it impossible to run on a AMD card? Else that statement makes no sense. Can i say the same about physX yes it's easy " PhysX will work on a 780 regardless of what AMD does."

The question was more: If you run a R9 290 AND a 780 in the same machine, will you be able to use PhysX (without driver hacks, etc)?
 
Apparently NVIDIA worked a lot closer with Valve to make SteamOS happen too. Now I'm sure one day benchmarks will show that NVIDIA hardware performs better on SteamOS than comparable AMD hardware... and no doubt people will complain that this is all because of some unfair advantage that NVIDIA coerced out of Valve. When the reality is that NVIDIA decided to show up while AMD was twiddling their thumbs. You know, the same reason why Intel stomps AMD in the CPU space.

And you think AMD hasn't been working with Valve? It's because currently AMD's drivers are not anywhere near as good as their Windows drivers. So for now it's not officially supported on SteamOS, but it's tested and working perfectly as well as Intel graphics.

What if I told you that open source drivers are where the action is? And Nvidia is failing horribly with open source support, where as AMD is doing a decent job. Right now open source drivers are superior on AMD then on Nvidia.

Don't be shocked if AMD decides to fully support open source drivers instead of their proprietary drivers, when it comes to Linux.
 
From what I've seen over the past few years, and what I've heard from developers who choose to come out and say it, is that AMD's engineering help pales in comparison to NVIDIA's. Sometimes they are lucky to even get a call returned from AMD while NVIDIA sends an engineer out the next day cross-country to the developers studio, etc.

Sounds like the game studios just want NVIDIA's quality of support yet also want NVIDIA to optimize their game for NVIDIA's competitor as well. Screw that.

Please feel free to list the names of the studios these developers work for.

Considering you dont seem to understand that the problem is that Nvidia is making it as hard as possible for AMD to optimize gameworks games, you should probably not keep spewing bull shit.
 
More of the same corporate protectionist bullshit from Nvidia. Proud to be an anti-[Nvidia Fanboy].
 
Ever heard of the AMD-64 instruction set? Ever heard of the Havok Physics engine?
the x64 instruction set was licensed to intel using the broad cross-license agreement the 2 companies have.

Havoc physics is an expensive commercial product, not an open standard. Btw, intel purchased havok a few years ago. :p
 
Continuing to say something that is wrong doesn't make it right.

AMD_Mantle_Is_Proprietary_Now_But_It_Will_Become_Widely_Available_to_Others

It's not at all the same situation as GameWorks, G-Sync, PhysX, etc.

Nice link. :p. I think it says "Mantle is proprietary now" lol

AMD cries because none of its proprietary initiatives ever succeeded. And you can't complain about nvidia having a developer program when AMD has one too, however less successful it is.
 
Nice link. :p. I think it says "Mantle is proprietary now" lol

AMD cries because none of its proprietary initiatives ever succeeded. And you can't complain about nvidia having a developer program when AMD has one too, however less successful it is.

Having a developer program is not the same as purposely trying to hobble the competition.
 
Having a developer program is not the same as purposely trying to hobble the competition.
lol, both AMD and Nvidia have programs that help developers optimize games for their own products. It's no worse when Nvidia helps developers optimize games for its products than when AMD does the same.

This thread delivers supreme silliness, and a lot of butthurt.
 
So instead of making their GPUs faster, writing better drivers, making games more optimised for their GPUs, NVIDIA has decided to spend money making sure games run slower on the competitions GPUs.

This benefits nobody, it only hurts gamers.

The money you spend on NVIDIA GPUs is not being used to make better GPUs but instead being used to make games slower.
 
lol, both AMD and Nvidia have programs that help developers optimize games for their own products. It's no worse when Nvidia helps developers optimize games for its products than when AMD does the same.

This thread delivers supreme silliness, and a lot of butthurt.

That is not like this, AMD or NVIDA will get access to games code early to develop optimised drivers for it.

After the game is out the other company can see the code and develop optimised drivers for the game.

In this case NVIDIA gets early access to the code to develop optimised drivers, but even after the game it out AMD is denied the ability to optimise drivers.
 
That is not like this, AMD or NVIDA will get access to games code early to develop optimised drivers for it.
That's part of the whole general development process, and as we see, optimized drivers aren't usually available for a release or so after a new game is released.

However, that's different from the logo branded optimizations both Nvidia and AMD do using their developer relations departments (which assist developers to optimize or add exclusive effects... AMD's last high profile title was Bioshock Infinite, until that Mantle Frostbite 3 patch is released). It's obvious, but Nvidia just has far more titles, which seems to be a sore point for AMD fans. Don't worry, it's not for lack of AMD trying! :p

The issue really is that Nvidia made a set of libraries and other developer support that makes implementing certain graphic engine features 1) easier since it's in a pre-made library and 2) optimized for speed. Nothing is stopping AMD from offering a similar product to developers. Just like nothing is stopping Nvidia from releasing a competitor to Mantle. The difference is that Nvidia users don't cry as much about not having Mantle. Who cares really? Mantle will likely flop like every single other AMD proprietary offering. :D
 
That's part of the whole general development process, and as we see, optimized drivers aren't usually available for a release or so after a new game is released.

However, that's different from the logo branded optimizations both Nvidia and AMD do using their developer relations departments (which assist developers to optimize or add exclusive effects... AMD's last high profile title was Bioshock Infinite, until that Mantle Frostbite 3 patch is released). It's obvious, but Nvidia just has far more titles, which seems to be a sore point for AMD fans. Don't worry, it's not for lack of AMD trying! :p

The issue really is that Nvidia made a set of libraries and other developer support that makes implementing certain graphic engine features 1) easier since it's in a pre-made library and 2) optimized for speed. Nothing is stopping AMD from offering a similar product to developers. Just like nothing is stopping Nvidia from releasing a competitor to Mantle. The difference is that Nvidia users don't cry as much about not having Mantle. Who cares really? Mantle will likely flop like every single other AMD proprietary offering. :D

You clearly have no clue what is going on and how it will hurt everyone, your bias is showing.

NVIDIA is deliberately preventing AMD from optimising, NVIDIA GPU users don't benefit from this so, and AMD GPU users suffer, so why are you supporting this.

In fact NVIDIA GPU users suffer as well, as this money is not being spend optimising for NVIDIA GPUs for making better GPUs.

So both NVIDIA and AMD users suffer.
 
You clearly have no clue what is going on and how it will hurt everyone, your bias is showing.

NVIDIA is deliberately preventing AMD from optimising, NVIDIA GPU users don't benefit from this so, and AMD GPU users suffer, so why are you supporting this.

In fact NVIDIA GPU users suffer as well, as this money is not being spend optimising for NVIDIA GPUs for making better GPUs.

So both NVIDIA and AMD users suffer.
The article speculates that nvidia is partly preventing amd from running optimizations at the shader level. IE shaders hacks for specific games, really amd shouldn't be relying on that too much for game fps, as that's a IQ issue but i guess AMD users are used to spotty IQ so they can claim more fps. It's a superfluous article designed as click and troll bait. AMD is troubled by a locked dev kit nvidia is offering as they can't inject code into that, instead they would have to offer developers a similar package which they don't want to do as it costs money. Developers aren't complaining AMD is complaining at best.
 
You clearly have no clue what is going on and how it will hurt everyone, your bias is showing.

NVIDIA is deliberately preventing AMD from optimising, NVIDIA GPU users don't benefit from this so, and AMD GPU users suffer, so why are you supporting this.
lol @ complaint. I don't really care what Nvidia does since I haven't bought an Nvidia card in over 2 years.

You are being one sided in your complaint: AMD works with developers to optimize games too, but they're just not as good at it as Nvidia is. Nvidia has to do sleuthing in those cases to see what is done to engineer shader replacements if necessary, and AMD could do the same if it wishes. The problem for AMD is that GameWorks may make the job a bit harder since many developers may use the libraries, and different revisions of those, in multiple ways.

And lol @ only Nvidia hurting gamers in some unspecified ways. AMD's proprietary Mantle is clumsily cleaving the gaming market on both AMD GPUs and Nvidia cards. AMD users are just going to cry more if GameWorks is extended to offer more Mantle-like low level support because Nvidia simply has a larger and better developer relations department, and more games optimized for Nvidia. Both things are almost a certainty. Think about that while you're pining high hopes on Mantle, because it's the same exact mess in the other direction. OMG how will poor widdle Nvidia figure out what proprietary Mantle is doing and how many Nvidia users will flood message boards with tears. lol
 
Ever heard of the AMD-64 instruction set? Ever heard of the Havok Physics engine?

AMD-64 instruction set is available to Intel because of the x86 cross-licensing deal that has been in place for over a decade now. Havok is middleware.
 
I'm just waiting for more Mantle fanboys to jump up and down chanting Mantle inside their sacrificial hymns to AMD.

Let's talk about Mantle when it is released... For all we know AMD could simply be inflating their product, noone here knows, just pure speculation.

Well, you know except for the developer demos showing off the power of mantle via massive space battles and developers stating their draw calls maxed out about ~15k/...minute? second? and Mantle was giving them 100k and they expect as hardware impoves over the next few years to reach 1 million draw calls by 2k18 vs ~15k via DirectX.
 
Well, you know except for the developer demos showing off the power of mantle via massive space battles and developers stating their draw calls maxed out about ~15k/...minute? second? and Mantle was giving them 100k and they expect as hardware impoves over the next few years to reach 1 million draw calls by 2k18 vs ~15k via DirectX.

Im hoping for the best, but I'll wait for release to see the final product.
 
Mantle was offered to everybody, but only AMD showed interest, being the nice Canadians they are. Maybe it's time for a lawsuit in this field?
 
Ah...nothing like AMD vs. nVidia thread to bring the best out of people. Kids...grow up.
 
ITT: nVidia is big and evil. AMD is perfect and if they do the same as everyone else it's their right.
 
the x64 instruction set was licensed to intel using the broad cross-license agreement the 2 companies have.

Havoc physics is an expensive commercial product, not an open standard. Btw, intel purchased havok a few years ago. :p

I don't want to argue about x86-64 but there's no evidence that HAVOK is expensive to license. Also, unless Intel made it worse, which I doubt very much, it should be working great on AMD CPUs. At least it used to.

These threads are always full with fanboys. I like nvidia products better but they have pulled a lot more shit than AMD over the years.
 
I don't want to argue about x86-64 but there's no evidence that HAVOK is expensive to license. Also, unless Intel made it worse, which I doubt very much, it should be working great on AMD CPUs. At least it used to.
Havok Physics has 5-6 digit license fee, starting around $70,000 per title (according to developers who have licensed it) if the title sells for over $10. Terms, but not license fees, are mentioned here: http://www.havok.com/try-havok

There has never been a "problem" running physics on the CPU, whether it's a custom designed physics engine or the array or free or commercial offerings, including PhysX. That's how it's generally been done since the dawn of computer gaming. That way is just far less scalable than GPU physics.

It's been over 3 years (mid-2011) since Nvidia updated the crufty Ageia codebase to move CPU physics from x87 to SIMD. 2010 called and wants its outrage back. ;)
 
Back
Top