delete this because less than half the people around here understand or even want to

[xfia]

https://www.youtube.com/watch?t=44&v=v3dUhep0rBs

 

[MxPhenom 216]

Why do you make these threads and post AMD slideshow shots?

 

[dorsetknob]

Do not feed the TROLL

 

[xfia]

because some people just go on ignoring what the people that actually make the stuff say.. highly cynical

 

[RejZoR]

Technically speaking, AMD had very advanced GPU's and they still are very advanced with certain goodies NVIDIA just started using. But how will this reflect THIS very moment, no one really knows until real DX12 games show up. But we need to first get Windows 10 so there's that...

 

[xfia]

Technically speaking, AMD had very advanced GPU's and they still are very advanced with certain goodies NVIDIA just started using. But how will this reflect THIS very moment, no one really knows until real DX12 games show up. But we need to first get Windows 10 so there's that...

http://www.guru3d.com/articles_pages/msi_radeon_r9_390x_gaming_8g_oc_review,6.html

 

[erocker]

Until people actually start using DX12, it's all theory-crafting.

 

[Aquinus]

Isn't this the third or fourth thread I've seen xfia post about AMD in the last 2 days? I think someone is too anxious for a review to calm down and wait.

 

[AsRock]

I don't believe it will, it may for a little while but nVidia will put HBM memory on their cards sooner or later which will give the boost what AMD have gotten.

Maybe if AMD have some thing to really improve their design before or when nVidia add HBM but we will have to wait and see.

Either way if we get 14nm their be boost all round, but as we know AMD and nVidia have no control over that.

 

[MxPhenom 216]

I don't believe it will, it may for a little while but nVidia will put HBM memory on their cards sooner or later which will give the boost what AMD have gotten.

Maybe if AMD have some thing to really improve their design before or when nVidia add HBM but we will have to wait and see.

Either way if we get 14nm their be boost all round, but as we know AMD and nVidia have no control over that.

I dont think its the HBM memory. Its GPU chip architecture.

 

[AsRock]

I dont think its the HBM memory. Its GPU chip architecture.

Thing is i bet nVidia have a pretty good plan for the next design, and lets face it all so has the money to try rush plans too if needed.

But yes your right it wont be all the HBM that's making the improvements.

This HBM will help AMD for the 1st 6 or so months even more so with those who want compact computers but nVidia will get their hands on it and dump tons of money in to it.

 

[MxPhenom 216]

Thing is i bet nVidia have a pretty good plan for the next design, and lets face it all so has the money to try rush plans too if needed.

But yes your right it wont be all the HBM that's making the improvements.

This HBM will help AMD for the 1st 6 or so months even more so with those who want compact computers but nVidia will get their hands on it and dump tons of money in to it.

I think Nvidia has had there hands on it ever since they announced the volta/pascal generation of GPUs (Nvidia CEO had the GPU module in his hand with the HBM memory and all). They were the first ones to talk about it in their conferences, but then AMD beat them to actually having a GPU ready with it implemented.

 

[AsRock]

I think Nvidia has had there hands on it ever since they announced the volta/pascal generation of GPUs (Nvidia CEO had the GPU module in his hand with the HBM memory and all). They were the first ones to talk about it in their conferences, but then AMD beat them to actually having a GPU ready with it implemented.

Totally not surprised all so thinking their after more than 4GB.

 

[LightningJR]

Correct me if I am wrong but doesn't a higher draw call only allow for more stuff to be on screen? If that is the case then the bottleneck will most likely be something else other than the draw calls and not giving any advantage to either AMD or NVidia.

The way I understand it is if you want to take advantage of the massive increase to draw calls in DX12 you need to have less detail overall to keep the GPU from struggling to render all the stuff on screen. I would bet that's why we see an RTS space sim as a draw call benchmark. The 3DMark draw call benchmark has very little geometry or texture and another DX12 game I found was a ground based RTS.

Plus don't most game creators make games based on the lowest common denominator?

If I am right the improved draw calls DX12 brings creates the potential for better looking games but unless a game is bottlenecked by by the draw call/s then I don't expect AMD to get any "unlocked" power and leave NVidia in the dust.

But I could be wrong. It did happen once before

 

[rtwjunkie]

Isn't this the third or fourth thread I've seen xfia post about AMD in the last 2 days? I think someone is too anxious for a review to calm down and wait.

Xfia lives, breathes and has AMD pumped in intraveneously.

 

[Frag_Maniac]

Going forward, I think the creators of Pci Express need to seriously consider abandoning it instead of updating it, for a faster interlink between GPU and CPU, like Nvidia is trying to do with NVLink. It really needs to happen by someone OTHER than a GPU designer to make it unilateral though.

IMB is the single biggest benefactor of NVlink of the four founders of Pci Express. If that kind of thing is to go mainstream on consumer MBs though, it really needs to be backed by both Intel and AMD, the latter of which is unlikely with Nvidia at the helm.

So right now I'd like to see Intel and IMB get together and come up with a successor to Pci Ex. If that happens I'm sure Dell and HP will jump on board.

As for draw calls and bottlenecks. The lower level API via directly sinking with GPU architecture made to utilize it allows more draw calls by the GPU, but there will also be less demand on the CPU due to consolidation of instruction sets. MS back with their Forza Dx12 demo explained they will now be able to cache repeat instruction sets, vs issue them over and over like before.

In laymen's terms, what used to be almost entirely manually issued instructions will be heavily automated with Dx12.

 

[revanchrist]

I think Nvidia has had there hands on it ever since they announced the volta/pascal generation of GPUs (Nvidia CEO had the GPU module in his hand with the HBM memory and all). They were the first ones to talk about it in their conferences, but then AMD beat them to actually having a GPU ready with it implemented.

You speak as if Nvidia invented HBM while in truth it was AMD who came out with the idea of HBM and they partnered with SK Hynix to develop HBM technology since 2010. HBM is the fruit of a 5-year long of investment from AMD. True, SK Hynix is selling HBM memory to Nvidia for their Pascal generation of GPUs but that doesn't make Nvidia the "pioneer" there.

 

[RejZoR]

Hm, does AMD have any licensing/patents involved in it from the Hynix partnership in development of this tech? Would be smart to have it, in the end, they'd be earning a little from every HBM chip sold to 3rd parties...

 

[btarunr]

Yes it will. And it will give AMD CPUs a nice advantage. That's why Pascal will feature simpler CUDA cores than Maxwell, and will have higher degree of parallelism overall. DX12's biggest feature is async shader and native multi-GPU. So you can pair a Fury X with a Titan X in multi-GPU and troll fanboys.

 

[Mussels]

Why oh why does this thread have such a flamebate title and options?
Why not "do you think AMD will gain a big advantage" or "Will this help AMD close the performance gap?"

Yeah, seeing the results from things like star swarm and 3Dmark make me think AMD has been planning this a long time.

http://i.imgur.com/20ytdvO.png


Those are someone elses results, but very similar to mine (Due to a bug with 3Dmark i cant test DX12 on this build of W10)
Long story short: Nvidia has better multi threaded DX11. in DX12, AMD seem to be skyrocketing ahead, and guess which company sells CPU's with lower single threaded clock speeds, but lots of cores?

 

[RejZoR]

VIA ? AMD certainly has more experience with async shaders, they've been using them for way longer than NVIDIA (or anyone else).

 

[mirakul]

No surprise if Dx12 unlock amd's performance. DX12 is literately Microsoft's Mantle by the way.
But I doubt that they can leave nVidia in dust

 

[Mussels]

No surprise if Dx12 unlock amd's performance. DX12 is literately Microsoft's Mantle by the way.
But I doubt that they can leave nVidia in dust

in draw calls at least, AMD have gone from a decently sized deficit and suddenly leapt ahead quite a lot. How that translates to real world performance, will need to be seen. In CPU intensive titles (RTS games) it could be a big thing.

 

[john_]

3D Mark's API feature test proved that AMD's drivers under DX11 DO have problems with draw calls. Because of that AMD's hardware is really getting a huge boost with DX12, and even performs better than Nvidia hardware. That's the reason Nvidia is preparing to defent it's possition through GameWorks and other software libraries that will try to put a few breaks on AMD's GPUs.

That thing alone is enough to consider that AMD will get a nice boost from DX 12 and probably will come on top in performance. Will they leave Nvidia in the dust? That depends on how much successful will be GameWorks in the future, and Nvidia does have plenty of money to make it successful.

PS One more thing to consider is that Nvidia's architecture needs constant optimizations, while GCN is something you can leave on auto pilot. That's why, in my opinion, AMD's cards that perform equally or a little worst compared to Nvidia models in the beginning, 2-3 years latter perform much faster compared to the same Nvidia models. So, by the time Pascal comes out, AMD cards will look better compared to today.

They were the first ones to talk about it in their conferences, but then AMD beat them to actually having a GPU ready with it implemented.

AMD co develop HBM with Hynix. Nvidia never had their hands on HBM before AMD. This phrase reminds me about DX12 and how Nvidia had access to it before AMD, but still the poor little AMD manages to come out with Mantle before we knew about the existence of DX 12.

High Bandwidth Memory - Wikipedia, the free encyclopedia

High Bandwidth Memory was proposed as an industry standard by AMD and Hynix, started in 2010 and adopted by JEDEC as JESD235 in October 2013.

Plus don't most game creators make games based on the lowest common denominator?

You mean PhysX and GameWorks?

 

[xfia]

in draw calls at least, AMD have gone from a decently sized deficit and suddenly leapt ahead quite a lot. How that translates to real world performance, will need to be seen. In CPU intensive titles (RTS games) it could be a big thing.

your failing to grasp what async shaders do and what they mean.. what is suppose to run on the cpu will be and what is suppose to be run on the gpu will be.. microsoft already said one gcn performance optimization was over 30 percent for lighting effects.
some people have called gpu's stupid throughout history well here it is the gpu brains for everyone that understands the async highway
nvidia may try push harder with gameworks but they will fail to the future.
dont be afraid to watch the animated videos! you get to learn and they make it simple so your kids can learn like its the discovery channel!