Monday, July 3rd 2017

NVIDIA Adds DirectX 12 Support to GeForce "Fermi" Architecture

With its latest GeForce 384 series graphics drivers, NVIDIA quietly added DirectX 12 API support for GPUs based on its "Fermi" architecture, as discovered by keen-eyed users on the Guru3D Forums. These include the GeForce 400-series and 500-series graphics cards. The support appears to be sufficient to run today's Direct3D feature-level 12_0 games or applications, and completes WDDM 2.2 compliance for GeForce "Fermi" graphics cards on Windows 10 Creators Update (version 1703), which could be NVIDIA's motivation for extending DirectX 12 support to these 5+ year old chips. Whether they meet your games' minimum system requirements is an entirely different matter.

Source: Guru3D Forums
Add your own comment

57 Comments on NVIDIA Adds DirectX 12 Support to GeForce "Fermi" Architecture

#1
Chaitanya


Still funny when I read Dx12 and nVidia together.
Posted on Reply
#3
plåtburken
Which version exactly of the drivers are we talking about?
Also does this apply to laptop cards or is it desktop only?
Posted on Reply
#6
Liviu Cojocaru
This is really strange but a nice thing I would say even after so long
Posted on Reply
#7
KainXS
I thought they gave up trying this years ago:wtf:
Posted on Reply
#8
qubit
Overclocked quantum bit
So Fermi gets DX12 support only two years after W10 came out? Impressive. /s
Posted on Reply
#9
bug
GoldenX said:
https://www.techpowerup.com/download/nvidia-geforce-graphics-drivers/ As long as the notebook GPU is Fermi.

Freaking finally, would be great to have Vulkan too, but they already failed on that promise.
What do you mean "finally"? Afaik, both Fermi and GCN 1.0 were left without support initially. Was GCN 1.0 support added a long time ago and I missed it?

Also, when it comes to Vulkan, it seems to me support is pretty much on par between AMD and Nvidia, so I'm not sure what's the failure here: https://en.wikipedia.org/wiki/Vulkan_(API)#Compatibility
Posted on Reply
#10
RejZoR
All GCN cards (HD7000 and above) had D3D_12_0 support from day 1 basically.
Posted on Reply
#11
Kaotik
It's not 12_0 capable, it's 11_0 capable. The only thing that changed really is the fact that now it can run the API itself.
(for reference, even Kepler and Maxwell gen 1. are 11_0, not even 11_1 like GCN1 & Haswell/Broadwell, let alone 12_0 like GCN2-4 or 12_1 like GCN5 and Maxwell gen 2. & Pascal and Skylake/Kaby Lake)

edit:
RejZoR said:
All GCN cards (HD7000 and above) had D3D_12_0 support from day 1 basically.
No they didn't. GCN1 is 11_1, 2-4 are 12_0, 5 is 12_1.
D3D12 supports feature levels 11_0, 11_1, 12_0, 12_1
Posted on Reply
#12
jabbadap
RejZoR said:
All GCN cards (HD7000 and above) had D3D_12_0 support from day 1 basically.
That is true and nice but Fermi is from hd5870 era.

Just wonder is those fermi cards any good for dx12? At least timespy score for gtx570 looks very low.
Posted on Reply
#13
Kaotik
jabbadap said:
That is true and nice but Fermi is from hd5870 era.

Just wonder is those fermi cards any good for dx12? At least timespy score for gtx570 looks very low.
See above.

D3D12 has 4 feature levels, 11_0, 11_1, 12_0 and 12_1.
Fermi, Kepler and Maxwell gen 1 are 11_0
GCN1 and Haswell/Broadwell are 11_1
GCN2-4 are 12_0
GCN5, Maxwell gen 2, Pascal and Skylake/Kaby Lake are 12_1

In addition to the feature level support, they can support variety of different Tiers on specific features.
As it stands today, GCN5 and Skylake/Kaby Lake have the most comprehensive support for D3D12 and it's FL12_1. (it is possible that GCN5 could actually surpass the Intels, but it all comes down to minimum precision - if it's FP16, they're the same, if it's FP10, it's the widest possible support you can have)
Posted on Reply
#14
bug
jabbadap said:
That is true and nice but Fermi is from hd5870 era.

Just wonder is those fermi cards any good for dx12? At least timespy score for gtx570 looks very low.
I don't think it's about actually using DX12 features as much as running DX12 code that can (natively) degrade gracefully.
Posted on Reply
#15
nemesis.ie
I think this needs to be called "fine beer". ;)
Posted on Reply
#16
rtwjunkie
PC Gaming Enthusiast
Kaotik said:
See above.

D3D12 has 4 feature levels, 11_0, 11_1, 12_0 and 12_1.
Fermi, Kepler and Maxwell gen 1 are 11_0
GCN1 and Haswell/Broadwell are 11_1
GCN2-4 are 12_0
GCN5, Maxwell gen 2, Pascal and Skylake/Kaby Lake are 12_1

In addition to the feature level support, they can support variety of different Tiers on specific features.
As it stands today, GCN5 and Skylake/Kaby Lake have the most comprehensive support for D3D12 and it's FL12_1. (it is possible that GCN5 could actually surpass the Intels, but it all comes down to minimum precision - if it's FP16, they're the same, if it's FP10, it's the widest possible support you can have)
What are you calling Maxwell gen 1? Just the 750 and 750Ti?

On topic, this quiet Fermi rollout mystifies me. Why would Nvidia not tell Fermi owners? It makes no sense to release the capability and not publicize it.
Posted on Reply
#18
EntropyZ
I'm just running a GTX 460 1GB (I can OC it to GTX 470 territory, but who cares amirite?) and I really don't care, since most DX12 titles won't even be able to run at 30fps on low, unless you drop the resolution to stupid levels.

Why bother? Just so you can say you delivered on a promise that the majority of people forgot about? (Edit: So this was just put in without boasting about it, weird.)

Kudos to them but I'd just prefer better Volta drivers instead, thanks. It's nice but, whyyyyyyyy lol.
Posted on Reply
#20
Sempron Guy
nemesis.ie said:
I think this needs to be called "fine beer". ;)
fine furnace would be more appropriate :D
Posted on Reply
#21
bug
EntropyZ said:
I'm just running a GTX 460 1GB (I can OC it to GTX 470 territory, but who cares amirite?) and I really don't care, since most DX12 titles won't even be able to run at 30fps on low, unless you drop the resolution to stupid levels.

Why bother? Just so you can say you delivered on a promise that the majority of people forgot about? (Edit: So this was just put in without boasting about it, weird.)

Kudos to them but I'd just prefer better Volta drivers instead, thanks. It's nice but, whyyyyyyyy lol.
As I posted above, DX12 code can and will degrade gracefully based on the feature level that is actually supported.
With this change, your card (I had one of those, too) will be able to run DX12 code, even if it will only use DX11 features. It's not a game changer, but a nice to have feature imho.
Posted on Reply
#22
Kaotik
rtwjunkie said:
I'm just trying to ensure we are all talking the same, not double checking him.
Yes, Maxwell gen 1. refers to GM107 aka 750/750 Ti
Posted on Reply
#23
jabbadap
Kaotik said:
Yes, Maxwell gen 1. refers to GM107 aka 750/750 Ti
Well mobile gm108 is maxwell gen 1 too, but for desktop graphics only gm107 derivations were made.
Posted on Reply
#24
lZKoce
Chaitanya said:


Still funny when I read Dx12 and nVidia together.
First thing that came to my mind, but Intel's one beats it by FAAR :D :

Posted on Reply
#25
Kaotik
jabbadap said:
Well mobile gm108 is maxwell gen 1 too, but for desktop graphics only gm107 derivations were made.
True, forgot about the GM108
Posted on Reply
Add your own comment