Discussing GCN Architecture, Performance & The Future

[melmac]

I know I'm likely right, I'll happily admit when I'm not.

I'll just have to wait till the tabloids post it so you can quote it on here for your own benefit




If you honestly believe that, that's hilarious.

I am not sure you are right, AMD have gone out of their way to assure people that there will be multivendor support. I think there will be, but, I don't think Nvidia will ever use it, unless it's a case of Mantle been so successful that they will have to use it.

Think of it just from an AMD point of view, if they are really want Mantle to succeed then they have to make it workable on many different types of GPU because if they don't then Mantle will only last as long as GCN lasts, which would be pretty stupid.

If you honestly believe that, that's hilarious.

I agree with you, I don't think Nvidia is going anywhere and I don't think they will fail because of Mantle, but, if people back in 1997 predicted that 3DFX would fail in a few years because they didn't take up directx , they would have been laughed at too.

I dunno, there's a difference between ranting and trying to point out the bleeding obvious.

It's only your opinion though, it's not the "bleeding Obvious" Unless you can give me evidence that proves that Mantle will never be open and Nvidia will never use it.

 

[ashmanuk67]

In three to five years time we will all look back and laugh at this or any other thread relating to MANTEL.

Mantel will be the norm all software and graphics card vendors will have a Mantel option in the rendering type and PC's gaming will be as popular with all types of game developers again

Mantel should be neutral between both AMD and Nvidia well that's what i would wish for

 

[Silent_Scone]

I am not sure you are right, AMD have gone out of their way to assure people that there will be multivendor support. I think there will be, but, I don't think Nvidia will ever use it, unless it's a case of Mantle been so successful that they will have to use it.

Think of it just from an AMD point of view, if they are really want Mantle to succeed then they have to make it workable on many different types of GPU because if they don't then Mantle will only last as long as GCN lasts, which would be pretty stupid.

I've said all along I'm sure it will have multi vendor support, but it's not going to be as clean as AMD are making out. NV won't adopt it, it's too much work for them. In fact arguably it's in their interest not to. It's in mine and yours as a consumer, but not Nvidias. I guess I'm a little sceptical of Mantles success, as everyone should be who's be around long enough. But for NV, they have control of Open Graphics Libs which they can amend freely. So you'll have to excuse me if I think Mantle is a bit of a stretch at the moment . Ask me again in 2015.

 

[ubersonic]

You're vastly overestimating Mantles success.

I can understand your POV, but I'm not really overestimating it, just taking in the whole picture and factoring in what will happen IF it takes off (which with the time/money/planning AMD have invested is quite likely).

Here's the important points I have focused on for AMD's offensive:

1: Their CPUs are making good IPC gains VS Intel - which means AMD closing the gap on Intel
2: Games/Applications are supporting multi core CPUs better every day - which again means AMD closing the gap on Intel
3: A game running mantle with an AMD GPU will run much better than it will running Direct3D on an Nvidia GPU of the same price - As the is no AMD-Nvida gap this means AMD pulling out in front (with software that supports mantle ofc, which will mean more and more as time progresses).

Now for the big gun:

4: An AMD APU will be able to offload CPU physics processing to the iGPU, freeing up CPU performance - two effects, firstly VS an Intel CPU it means the CPU can offload work the Intel CPU can't which again closes the gap already closed by IPC improvements and better core utilization (lets be honest, the is no gap now), secondly it means unlike an Nvidia GPU solution the GPU doesn't take a performance hit for doing physics (PhysX) work.

So to sum it all up, in a Mantle game with a high end AMD 6/8 core APU (I know they don't make one ATM) you would get much higher CPU and GPU performance than an Intel/Nvidia combination for much less cost, if that isn't enough to shake up the market I don't know what is.

 

[TheRealDeal]

I have to admit AMD's overall position looks to be getting stronger but Intel and Nvidia are better known and Bigger brands. It's going to take a great effort on amd's part to eat into there lead. I can see them eating into Nvidia's lead and they probably are as we speak. Intel on the other hand don't seem to lose anything quarter to quarter. Amd are said to be working on a 16 core cpu atm which could be pretty interesting. On the apu side if i was buying for my kids it would be an amd apu everyday of the week as they like to play minecraft.

 

[Silent_Scone]

You've come at it all wrong there. Yes it means high end performance is better achieved on otherwise slower 4 to 8 core AMD CPUS, yes through improvements in CF technology they're able to adapt tasks, yes it will be largely more efficient than Direct3D.

And in an ideal world this would be fantastic...I'm probably going to buy a 290X again just so I can see it for myself. But guess what, that doesn't matter. This is the most industry driven industry in existence - and two competing corporations are going to fall out if not now, then in a couple of years time. And there will be equal blame...It's as simple as this..."OpenGL is far more accessable, why should we (Nvidia) have to adopt this when you won't let us touch these core elements?"

AMD: "Well, we wrote it? And we don't want you knowing these certain aspects of GCN"

NV: "No thanks"

If Mantle is as good as it's being played up to be, I will be switching in due course believe me. But as an industry standard it's a long, long road and probably won't ever see it's destination as far as I'm concerned. You can throw all the publication at me you want. It's just not very likely, sorry

 

[TheRealDeal]

You've come at it all wrong there. Yes it means high end performance is better achieved on otherwise slower 4 to 8 core AMD CPUS, yes through improvements in CF technology they're able to adapt tasks, yes it will be largely more efficient than Direct3D.

And in an ideal world this would be fantastic...I'm probably going to buy a 290X again just so I can see it for myself. But guess what, that doesn't matter. This is the most industry driven industry in existence - and two competing corporations are going to fall out if not now, then in a couple of years time. And there will be equal blame...It's as simple as this..."OpenGL is far more accessable, why should we (Nvidia) have to adopt this when you won't let us touch these core elements?"

AMD: "Well, we wrote it? And we don't want you knowing these certain aspects of GCN"

NV: "No thanks"

If Mantle is as good as it's being played up to be, I will be switching in due course believe me. But as an industry standard it's a long, long road and probably won't ever see it's destination as far as I'm concerned. You can throw all the publication at me you want. It's just not very likely, sorry

It has a chance but like you say there is a long way to go for it to become an industry standard. I don't see microsoft lying down on this one. Amd have also said in so many words mantle is there to push the industry forward. With this in mind microsoft may have to rework dx to compete which is still pushing the industry forward.

 

[Silent_Scone]

Microsoft are at crossed paths because furthing DX could negatively impact XB1 sales, which currently no matter how low the potential is a financially poor idea as they'll make nothing to very little for a couple of years on it yet.

Basically yeah, MS are as much to blame as anyone

 

[TheRealDeal]

Microsoft are at crossed paths because furthing DX could negatively impact XB1 sales, which currently no matter how low the potential is a financially poor idea as they'll make nothing to very little for a couple of years on it yet.

Basically yeah, MS are as much to blame as anyone

Yep they have a really tough decision to make as it's not only mantle they have to worry about with steam os using Linux and OpenGl. I feel if amd keep a steady stream of big name developers coming in they might jump. We shall see how the steambox goes along with the new OS. I don't really think its gonna take off mainly due to If i want to play on a pad i have a ps4 with sony only games plus lots of cross platform games. Then i have my pc which i can also play on my tv with a pad if i wish but mouse and keyboard is still the best way to play most games.

 

[Silent_Scone]

Yep they have a really tough decision to make as it's not only mantle they have to worry about with steam os using Linux and OpenGl. I feel if amd keep a steady stream of big name developers coming in they might jump. We shall see how the steambox goes along with the new OS. I don't really think its gonna take off mainly due to If i want to play on a pad i have a ps4 with sony only games plus lots of cross platform games. Then i have my pc which i can also play on my tv with a pad if i wish but mouse and keyboard is still the best way to play most games.

Unfortunately most of the community doesn't agree with you there which is another contributing factor as to why this very thread exists in the first place lol

 

[TheRealDeal]

Unfortunately most of the community doesn't agree with you there which is another contributing factor as to why this very thread exists in the first place lol

Everyone has there opinions but from my experience most console gamers won't game on a pc where as pc players will happily play certain games on the console even though they consider it inferior. Console players in the main just don't see the benefits of playing on a pc and some will tell you they can't see the better visuals on the pc. I think the steam box is trying to target an audience which is very small and a lot of in between solutions fail or get massive. Its certainly a double edged sword.

I agree though mouse and keyboard is in my opinion the best interface to play most games (should have said that in my first post). The console gamers would certainly disagree. I used to be the same back in pre 2000.

 

[Rroff]

It was EA's decision to rush the game out though in an unfinished state and that was to damage cod sales.

To be fair i can't remember the last time it crashed for me, got to be over a month at least.

The game is broken at a far deeper level than just EA rushing it or crashing.

 

[Silent_Scone]

Would tend to agree, there is a lot of 'clipping' going on that they've yet to sort. The net code on the whole is tragically broken still compared to BF3

 

[Diagro]

I always lol when i see people complain about "code" when they prob don't have a clue how to fix it thyself

 

[GoogalyMoogaly]

Even if the performance of the Mantle API isn’t up to 45 percent per game, and only works out to be half as much, that’s still going to be a significant difference. This can turn a title barely running in 30FPS into the comfortable to play territory, or titles which are just shy of the target of 60 shooting way past that number.

A title barely running 30 FPS with a 22.5% increase will move into the 'comfortable to play territory'? It'd be 36.75FPS. I wouldn't find tis comfortable to play.
Just shy of 60 FPS, so say 59fps can shoot way past this number?
It'd be 72 fps. While this is definately a nice chunk above 60 FPS I'm not sure i'd consider it 'shooting way past that number'.

Both of these are a personal preference thing though I guess.
I'm not sure it'll sway the market quite as much as you'd think from the hype. I think AMD's pricing will do as much to affect sales as the performance increase.
Largely you're either at the low FPS where the increase won't help or you're at high enough FPS where you don't really need the extra.

45% is another story though...

 

[ubersonic]

AMD: "Well, we wrote it? And we don't want you knowing these certain aspects of GCN"

NV: "No thanks"

Ironically that's exactly how it went when Nvidia offered AMD PhysX

I always lol when i see people complain about "code" when they prob don't have a clue how to fix it thyself

I would have fixed it pre-emptivly: by not releasing the game until it was working/finished

 

[Silent_Scone]

I always lol when i see people complain about "code" when they prob don't have a clue how to fix it thyself

I lol more when someone kills me when I've been behind a concrete post for over 2 seconds. I've got some idea of how to fix it, anyway. Sever side HD would be a start. That would remedy a lot of the delay, just some people would nag about the performance loss on older hardware lol.

Ironically that's exactly how it went when Nvidia offered AMD PhysX

Most likely. I don't think for an instant that NV would open up the core libs to them. Although PhysX is arguably something AMD can do just fine without, and is a speck on the landscape compared to your very own API. But that doesn't mean AMD are obligated to spread their legs open. You'd be very naive if you thought AMD would trust Nvidia by letting them amend or simply just see core aspects of Mantle, for what it would tell them about GCN.

 

[PhysicsMan94]

I think it has a very bright future as the current development has obviously been planned by AMD for a very long time, since before they even considered buying ATi, and with their red steamroller* building momentum they look set to knock Nvida/Intel/Microsoft for six unless they band together which is unlikely.

My predictions are that Nvidia will try and counter Mantle with something and fail then fall behind, have to embrace GCN or something but end up diminished, much like 3DFX in it's later years. Intel will not realise the trouble they are in until £500 Haswell-E CPU's are being matched in games by sub £200 AMD APU's and by the time they mount a counterattack (possible partnership with Nvidia?) it will be AMD's biggest rout since the Athlon XP. Microsoft being Microsoft will try and deal with mantle with FUD and Windows issues like they did with OpenGL, but gamers/devs today won't believe the FUD and the Windows issues will do more damage to Windows as a gaming platform than Mantle.

Are you out of your mind? You're making it sound like AMD are going to take over the world, or the entire technology industry at least...

Nvidia are not in trouble, there are a heck of a lot of smart guys working at Nvidia and they know what they are doing - you seem to be forgetting that they are not just a company which designs and develops gaming graphics cards.

Intel are most certainly not in trouble either, there's a plethora of reasons why everyone suggests Intel CPUs for systems with a budget higher than £500. Also, Intel are (as with everyone else it seems) moving into the mobile market with their x86 processors as Broadwell is apparently going to be implemented into tablets - such a move will allow Intel to cover the desktop CPU market and enter the mobile CPU market. Thus, they are in no danger - and I think that they are probably not worried in the slightest about AMDs APUs or Mantle. Also, they have a market capitalisation of ~$128.5 Bn versus AMD's ~$3.03 Bn.

And Microsoft, well, I don't even know what to say if you honestly think they are in danger from AMD. Again, MSFT has market capitalisation of ~$303.7 Bn; which is around 100 times larger than AMDs. They are not in trouble.

 

[Gregster]

I always lol when i see people complain about "code" when they prob don't have a clue how to fix it thyself

What a daft thing to say. I buy something and it doesn't work properly (let's say a GPU) and I know that it is the GPU at fault, do I need to know the ins and outs of said GPU before I am allowed to say it doesn't work as it should?

 

[Suarez7]

Careful with the personal attacks - w3bbo

If AMD get Mantle right it could change a few things but on the grand scheme I don't think it will affect nvidia at all, also they are way to stubborn to adopt Mantle as we all know.