• 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.

Nvidia GeForce GTX 980 Ti Coming This Summer Featuring 3072 CUDA Cores and 6GB of GDDR5

Status
Not open for further replies.
It's okay I already predicted that reply you made :)

I never once suggested they wouldn't be able to run games in a few years but it seems some here think they'll be running games as well as they do now in a few years because of DX12.
 
People will want to sell their 970s for 10nm cards. It won't be out of reach for most people as the buy-in will likely end up being <£200.

You won't see pooled VRAM implemented in any meaningful way before 10nm.
 
Ayahuasca said:
I never once suggested they wouldn't be able to run games in a few years but it seems some here think they'll be running games as well as they do now in a few years because of DX12.
Click to expand...

If DX12 does what it says on the tin, then its equivalent to jumping a generation of GPU's in terms of performance.

So 970's (and other maxwell) cards could well be playing games just as well as they do now.

Of course it's all relative and the new GPU's of 2017-2018 will still be preferable, but there might not be a need to junk older cards as quickly as we used to.
 
I hope that is the case but it would kill profits for AMD and nVidia if users didn't have to upgrade as often and whilst DX12 will improve performance it will also bring more advanced graphics that are going to require more grunt regardless.
 
Ayahuasca said:
I hope that is the case but it would kill profits for AMD and nVidia if users didn't have to upgrade as often and whilst DX12 will improve performance it will also bring more advanced graphics that are going to require more grunt regardless.
Click to expand...

Very true this. I am proper excited for DX12 and like most of us enthusiasts, when we struggle to get games running well, we will upgrade.
 
I might replace my 780 non ti sli setup with one of these and pick up a second card at a later date since the Titian X is faster than my current setup even in a single card config
 
Ayahuasca said:
I hope that is the case but it would kill profits for AMD and nVidia if users didn't have to upgrade as often and whilst DX12 will improve performance it will also bring more advanced graphics that are going to require more grunt regardless.
Click to expand...

I think to fully implement all the advantages of DX12 like stacked memory in multi GPU setups we are going to need some massive upgrades in hardware and GPU grunt. For example to use stacked memory we will be to be running PCI-E 4.0 X16/X16 as there is no way that a PCI-E 3.0 X16 setup will keep up with the workload.
 
LambChop said:
From a pcie 3 perspective you can get max 32gb/sec bandwidth

How much bandwidth does stacked memory need ?
Click to expand...

Using 4 cards at very high resolution really pushes PCI-E 3.0 doing it the old fashioned way with DX11.

If you are going to stack memory with DX12 under the same conditions it is going to involve a great deal more info going backwards and forwards across the bus.

To answer your question I don't think anyone will know the answer but remember you can stack memory with Mantle too and we all know how many games support that (zero).
 
LambChop said:
From a pcie 3 perspective you can get max 32gb/sec bandwidth

How much bandwidth does stacked memory need ?
Click to expand...

According to this even PCI-e 4 isn't going to be fast enough

"Coming to the final pillar then, we have a brand new feature being introduced for Pascal: NVLink.
NVLink, in a nutshell, is NVIDIA’s effort to supplant PCI-Express with a faster interconnect bus.
From the perspective of NVIDIA, who is looking at what it would take to allow compute workloads to better scale across multiple GPUs,
the 16GB/sec made available by PCI-Express 3.0 is hardly adequate.
Especially when compared to the 250GB/sec+ of memory bandwidth available within a single card.
PCIe 4.0 in turn will eventually bring higher bandwidth yet,
but this still is not enough. As such NVIDIA is pursuing their own bus to achieve the kind of bandwidth they desire."

source
 
whyscotty said:
According to this even PCI-e 4 isn't going to be fast enough

"Coming to the final pillar then, we have a brand new feature being introduced for Pascal: NVLink.
NVLink, in a nutshell, is NVIDIA’s effort to supplant PCI-Express with a faster interconnect bus.
From the perspective of NVIDIA, who is looking at what it would take to allow compute workloads to better scale across multiple GPUs,
the 16GB/sec made available by PCI-Express 3.0 is hardly adequate.
Especially when compared to the 250GB/sec+ of memory bandwidth available within a single card.
PCIe 4.0 in turn will eventually bring higher bandwidth yet,
but this still is not enough. As such NVIDIA is pursuing their own bus to achieve the kind of bandwidth they desire."

source
Click to expand...

Looks like we will have to wait for PCI-E 6.0 before we can stack memory.:D

Unfortunately we will be on DX15 by then.:D
 
Kaapstad said:
Looks like we will have to wait for PCI-E 6.0 before we can stack memory.:D

Unfortunately we will be on DX15 by then.:D
Click to expand...

:(
maybe with compression it can be done
hbm just getting lots cheaper might be the way tho

hope we get to see apu's with hbm too and seeing how good that is, some interesting stuff coming ^^;
 
KillBoY_UK said:
I might replace my 780 non ti sli setup with one of these and pick up a second card at a later date since the Titian X is faster than my current setup even in a single card config
Click to expand...

This is what I've decided to do. Should help with case temps and no more sli issues. If my 780's had been 6gb models id have skipped this generation.
 
Status
Not open for further replies.
Back
Top Bottom