Pretty much the same way that AMD just had to embrace PhysX then....Just saying
That's different and you know it.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Discussing GCN Architecture, Performance & The Future
- Thread starter LtMatt
- Start date
That's different and you know it.
Guess that's another one off the christmas card list. Gregorio, Whyscotty and Rusty Galloway are my only Nvidia friends on the forum.
Eh hmmm don't be grouping me based on my current GPU
.That's different and you know it.
At a low level of detail it is, like all things, completely different. But if you abstract a little Mantle is no more realistic for nVidia than PhysX ever was for AMD.
I really need to get thinking of a nickname for Whyscotty. For some reason its not so easy with him. Efforts must be doubled on my part.
Regarding Physx and Mantle, im pretty sure AMD won't charge a licensing fee to use Mantle. If AMD keep Mantle to just their own gpu's it won't work. It will become like Physx and will see a small number of titles each year and no more. Mantle has to become open and everyone knows it. A long term closed API is not going to get widely adopted.
Regarding Physx and Mantle, im pretty sure AMD won't charge a licensing fee to use Mantle. If AMD keep Mantle to just their own gpu's it won't work. It will become like Physx and will see a small number of titles each year and no more. Mantle has to become open and everyone knows it. A long term closed API is not going to get widely adopted.
Last edited:
From what i have read, They said that the vision is to have an open API, but at the moment that's not feasible.
It has to be GCN only for a while i think. Afterall mantle is not finished, its not 1.0 or 1.1, its beta. I'm sure AMD won't want to rush it to leverage an advantage for a while and who can blame them. However i think we all know it will have to become open to be long term successful. They'll want as many people using it as possible, especially if the performance increase reported is accurate.
That's different and you know it.
Well according to your God, Roy Taylor, he said they were waiting for AMD to get in touch but I guess on this occasion, he was talking rubbish because he isn't bigging up AMD but since joining AMD, everything he said is true and nothing is wrong.
Not that I care but we can all grab a statement and say "I like it, so it is true" and that seems to be the order of Mantle.
All good and I am certainly not arguing about it again and like I said earlier...Just saying DoorMatt
If AMD ever lock Nvidia out from Mantle i'll hold up my hands Gregorio. I think we both know that won't happen though as doing that will not make Mantle a success. If we look at open and proprietary features from both sides AMD have a history of being open with their tech whereas Nvidia do not. If this changes with Mantle long term then ill admit i was wrong. I don't see that happening though.
EDIT
I know you love Royston as much as i do so don't pretend otherwise. It would be great for all three of us to go down the pub the shoot the ****.
EDIT
Regarding DoorMatt you can walk all over me anytime, tuts.
Soldato
- Joined
- 5 Sep 2011
- Posts
- 12,883
- Location
- Surrey
Unless im mistaken, your argument is that Nvidia can't use Mantle? Well the quotes ive provided contradict that opinion completely. I posted them on the first page. Mantle is not tied to GCN and can work on competing GPU solutions. It will be up to Nvidia to embrace and support Mantle in the end. It will only work on GCN initially but its still in Beta and support for other gpu's will be added further down the line. That's the plan and currently i have no reason to doubt that information is false. Keeping Mantle to GCN only makes no sense.
Good man. Now lets hug this out.
I never said they couldn't. I said it's not as clean cut as they're making out. At its core it will contain a lot of information about GCN and in all honesty its difficult not to compare aspects to the GW arguement. That is the first thing I said. Which is when you threw the trusty multiquote of a 10 milion dollar investment at me, which was when I elaborated by explaining he is talking as any developer would in his situation. . . In his element. Johan knows Mantle can be adopted but he is saying it at his own discretion as, like I said before, he's at no more liberty to say so than I am. Ask him on twitter whether he thinks NV will adopt Mantle. I encourage you to
Last edited:
I never said they couldn't. I said it's not as clean cut as they're making out. That is the first thing I said. Which is when you threw the trusty multiquote of a 10 milion dollar investment at me, which was when I elaberated by explaining he is talking as any developer would in his situation. . . In his element. Johan knows Mantle can be adopted but he is saying it at his own discretion as, like I said before, he's at no more liberty to say so than I am. Ask him on twitter whether he thinks NV will adopt Mantle. I encourage you to
The problem is not 'can Nvidia support Mantle'. The problem is 'will Nvidia support Mantle'. I believe they will be able to (in the future) but they won't. Even if it means their users lose out on perhaps the best thing to happen to the gaming industry in recent times. Time will tell.
Pretty much the same way that AMD just had to embrace PhysX then....Just saying
AMD didn't go with PhysX because they went with Havok. And it was a logical choice for them because Intel were pushing CPU physics, and many people were saying at the time that PhysX wouldn't last.
So you had Intel/Havok on one side and Nvidia/PhysX on the other, As a primarily CPU business which would you have gone with?
That sounds like it is a case of AMD say "here is all the source code, now fill your boots nVidia peoples". There will probably be far more to it than that (so much small print), this is why nVidia will not be using it. I could be massively wrong though and AMD do just that, which would be nice.
Soldato
- Joined
- 5 Sep 2011
- Posts
- 12,883
- Location
- Surrey
They won't but as I said earlier it isn't only NV being NV.
Mantle is built, no matter what spin is floating around, to maximise the potential of GCN architecture. To say it's not means it's a compromise of its very purpose. AMD will have control of all the core libraries at both a technical and legal level. Are you telling me with that in mind they expect NV of all people to adopt it?
lollers
. Cant and won't and that is why the industry won't ever make big leaps.OpenGL is a true open source library that NV can amend at their own leasure. AMD havd intentially made Mantle as appealing to NV as a Royal to a p*** pan
Last edited:
Ermmmm without looking it up, wasn't AMD charging down the Bullet Physics route? Correct me if I am wrong though
Ermmmm without looking it up, wasn't AMD charging down the Bullet Physics route? Correct me if I am wrong though
Nope, well not at that stage.
They went with Havok first, bullet physics came later.Havok isn't exactly the same as the havok fx that ati used.
Nope, well not at that stage.
Yer, my memory is bad (blame my age) and couldn't remember what way round it was. I actually thought Bullet was dead before it was finished but looking around, it seems used (without looking in depth).
Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,619
- Location
- Cornwall
If Nvidia had any intention of taking up Mantle I doubt they'd be putting the effort into OpenGL that they are.
Also, Mantle being open to Nvidia and Mantle being a levelling playing field with regards to Nvidia hardware are different things. Maybe Nvidia could implement Mantle, but there might have to be a much heavier software layer for Nvidia, maybe in some cases them having to do a software solution for a certain aspect that AMD wrote with their GCN architecture very much in mind. I doubt there will be very much in there aim at letting CUDA cores stretch their legs.
So this could mean a much lower performance boost for Nvidia hardware. In this case Nvidia might be just as well sticking with DirectX (maybe DirectX 12 will lower the gap somewhat too) or OpenGL and in doing so make Mantle an API that can only be used on AMD cards. Thus meaning the industry has to support DirectX/OpenGL as well. This in itself could mean that some (smaller) developers may decide not to bother with it as it'll take too much time and effort for the small team and will only help maybe 50% of its customers. I believe even the great and infallible Johan said that it might not be suitable for all developers.
Nvidia could have course develop a rival API based on similar principles but that does play to their cards strengths. They could make it open, although not terribly well suited to anyone else's hardware and argue that AMD could adopt their API.
I'm guessing AMD will be more likely to implement Mantle on their non-GCN cards, but have they guaranteed the same 45% performance increase on these cards?
What if the increase on these cards was only 15%? What if the increase was 15% and it was Nvidia's API and Nvidia had sole control over it (along with WB Games and Ubisoft, maybe). Would it seem like such a good idea for AMD to join in with that? Instead of using an API by Microsoft, you could use a slightly better API controlled by one of your main rivals. How tempted do you think they'd be?
Or let's assume that when Mantle is made open that it give about the same boosts on Nvidia hardware. What about, if in the mean time, Nvidia counter not only with OpenGL, but also by pulling out all the stops with their hardware and producing something that through brute force can narrow the gap with Mantle to a small amount. How eager will AMD now be to let Nvidia use Mantle with all the benefits, turning an AMD performance advantage into an Nvidia one. In this scenario might AMD release Mantle once it's out of beta, or might they wait until their hardware is in a better place to compete. Why would Nvidia want to give AMD this option and this kind of power?
Also, Mantle being open to Nvidia and Mantle being a levelling playing field with regards to Nvidia hardware are different things. Maybe Nvidia could implement Mantle, but there might have to be a much heavier software layer for Nvidia, maybe in some cases them having to do a software solution for a certain aspect that AMD wrote with their GCN architecture very much in mind. I doubt there will be very much in there aim at letting CUDA cores stretch their legs.
So this could mean a much lower performance boost for Nvidia hardware. In this case Nvidia might be just as well sticking with DirectX (maybe DirectX 12 will lower the gap somewhat too) or OpenGL and in doing so make Mantle an API that can only be used on AMD cards. Thus meaning the industry has to support DirectX/OpenGL as well. This in itself could mean that some (smaller) developers may decide not to bother with it as it'll take too much time and effort for the small team and will only help maybe 50% of its customers. I believe even the great and infallible Johan said that it might not be suitable for all developers.
Nvidia could have course develop a rival API based on similar principles but that does play to their cards strengths. They could make it open, although not terribly well suited to anyone else's hardware and argue that AMD could adopt their API.
I'm guessing AMD will be more likely to implement Mantle on their non-GCN cards, but have they guaranteed the same 45% performance increase on these cards?
What if the increase on these cards was only 15%? What if the increase was 15% and it was Nvidia's API and Nvidia had sole control over it (along with WB Games and Ubisoft, maybe). Would it seem like such a good idea for AMD to join in with that? Instead of using an API by Microsoft, you could use a slightly better API controlled by one of your main rivals. How tempted do you think they'd be?
Or let's assume that when Mantle is made open that it give about the same boosts on Nvidia hardware. What about, if in the mean time, Nvidia counter not only with OpenGL, but also by pulling out all the stops with their hardware and producing something that through brute force can narrow the gap with Mantle to a small amount. How eager will AMD now be to let Nvidia use Mantle with all the benefits, turning an AMD performance advantage into an Nvidia one. In this scenario might AMD release Mantle once it's out of beta, or might they wait until their hardware is in a better place to compete. Why would Nvidia want to give AMD this option and this kind of power?
1. Mantle seems great from what the people that have seen what it can do
2. Mantle could be usable at some point in the futures possibly on all gpu's
Point 1 well hopefully the wait wont be too long until we can all see what dfifference it makes for ourselves
Point 2 AMD have stated its only on there GCN cards until its sorted and out of beta + whoever many more versions before they hypothetically make it open.
Point 3 in the foreseeable future mantle is not a option offered even to Nvidia let alone nvidia taking it up
2. Mantle could be usable at some point in the futures possibly on all gpu's
Point 1 well hopefully the wait wont be too long until we can all see what dfifference it makes for ourselves
Point 2 AMD have stated its only on there GCN cards until its sorted and out of beta + whoever many more versions before they hypothetically make it open.
Point 3 in the foreseeable future mantle is not a option offered even to Nvidia let alone nvidia taking it up