• 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.

CA has done the worst DX12 port in Total Warhammer

Doom112

Doom112

Doom112

Permabanned
Joined
8 Jul 2016
Posts
430
One of the biggest advantages of DX12 – at least according to Microsoft – is its improved multi-threading capabilities. However, as we can clearly see in both the in-game benchmark and our Extreme scenario, the game is unable to properly take advantage of all our CPU cores (again, DX11 graph is on the left whereas DX12 graph is on the right).
Click to expand...

http://www.dsogaming.com/news/report-total-war-warhammer-runs-27-slower-dx12-nvidias-hardware/

This game shows that Nvidia is far better in DX11 in this game compare AMD in DX12.

http://www.pcgameshardware.de/Total...Specials/Direct-X-12-Benchmarks-Test-1200551/

CA never worked with NVIDIA on DX12 or took their input, which, one of the reasons the results are terrible on DX12. If this goes on then DX12 will have the same faith as DX10 ,therefore, developers need to take this brand sponsorship out and work with both parties.

We’re pleased to confirm that Total War: WARHAMMER will also be DX12 compatible, and our graphics team has been working in close concert with AMD’s engineers on the implementation. This will be patched in a little after the game launches, but we’re really happy with the DX12 performance we’re seeing so far, so watch this space!
Click to expand...

http://wiki.totalwar.com/w/Optimisation_Blog


Once again sorry for my bad English.
 
Last edited:
Doom112 said:
http://www.dsogaming.com/news/report-total-war-warhammer-runs-27-slower-dx12-nvidias-hardware/

This game shows that Nvidia is far better in DX11 in this game compare AMD in DX12.

http://www.pcgameshardware.de/Total...Specials/Direct-X-12-Benchmarks-Test-1200551/

CA never worked with NVIDIA on DX12 or took their input, which, one of the reasons the results are terrible on DX12. If this goes on then DX12 will have the same faith as DX10 ,therefore, developers need to take this brand sponsorship out and work with both parties.



http://wiki.totalwar.com/w/Optimisation_Blog


Once again sorry for my bad English.
Click to expand...

Not likely to ever happen, Nvidia are normally the worst culprits for this behavior, having their game gimped in 1 game does not counter the effect of their Gameworks gimping many many more titles for AMD games.

I would like to personally see less bias towards either vendor in games and dev teams to tune the games equally for both, this will never ever ever happen though, Nvidia actively throw money at developers to include their Gameworks features, AMD less so, all the time this practice happens it will never be an even footing for either party.
 
It's because NVidia doesn't support Async Compute in hardware fully on Pascal, and not at all on previous GPU's, and by the looks of it, that's what this patch has implemented.

NVidia does have 'fast context switching', what ever that is it's not a full Async Compute implementation as in the DX12 standard.
 
hominid said:
It's because NVidia doesn't support Async Compute in hardware fully on Pascal, and not at all on previous GPU's, and by the looks of it, that's what this patch has implemented.

NVidia does have 'fast context switching', what ever that is it's not a full Async Compute implementation as in the DX12 standard.
Click to expand...

There is also this as well, A-Sync is not so good in Maxwell, Pascal does not seem to be so bad as Maxwell is.
 
You should think of Async Compute as AMD's Tessellation :D

AMD will leverage their hardwares ability to do it to hurt the opposition, just like NVidia did with Tessellation back when their cards were struggling. The only thing they could leverage was the good tessellation ability as it performed not so well on AMD - Crysis 2 anyone?
 
The main purpose of DX12 and its existence is to reduce CPU bottleneck not Async. Async is just a tiny feature of DX12 .


DX12 has failed to do its main feature.
 
Last edited:
hominid said:
It's because NVidia doesn't support Async Compute in hardware fully on Pascal, and not at all on previous GPU's, and by the looks of it, that's what this patch has implemented.

NVidia does have 'fast context switching', what ever that is it's not a full Async Compute implementation as in the DX12 standard.
Click to expand...

Did you look at the DX11 benchmark? Fury X with Async in DX12 is still 30% slower then GTX 980 Ti DX11.
 
Fury in comparison to a 980Ti is crap though, 980Ti is in a different league imo.

Async is an api optimisation feature that works best on a hardware level, heavy/over use of tessalation is a different ball game, again imo.

LoadsaMoney said:
Pascal has got to have good a-sync, or Nvidia wouldn't have had the a-sync support patched into ROTTR.
Click to expand...


Or perhaps Nv only got 'X' amount of time based exclusivity of not having Async and any other non vendor exclusive optimisations integrated into the title as AMD's pre patch mins were terrible and have since massively improved post Async patch?
 
SiDeards73 said:
Not likely to ever happen, Nvidia are normally the worst culprits for this behavior, having their game gimped in 1 game does not counter the effect of their Gameworks gimping many many more titles for AMD games.

I would like to personally see less bias towards either vendor in games and dev teams to tune the games equally for both, this will never ever ever happen though, Nvidia actively throw money at developers to include their Gameworks features, AMD less so, all the time this practice happens it will never be an even footing for either party.
Click to expand...

+1
 
Doom112 said:
Did you look at the DX11 benchmark? Fury X with Async in DX12 is still 30% slower then GTX 980 Ti DX11.
Click to expand...

It's a lot smoother though . Maintains minimums better and Dx12 mode for TW is still in beta.

The dso link was talked about in the TW thread. I couldn't reproduce the CPU issue they had.. I got much better CPU loading when using dx12.

As you can see from guru 3d's actual review.

http://www.guru3d.com/articles-page...-graphics-performance-benchmark-review,8.html
 
Fair comments from Doom112 and when you compare DX12 AoTS on a Fury X, to DX11/12 on a 980Ti, the DX12 implementation for AMD is very good and according to a few of the tech sites, AMD have the lead with their DX12 version. Not sure why there is any disagreement with this thread and DX11 on NVidia's 980Ti is showing 30% faster than DX12 on AMD in Warhammer. Some of you guys are letting bias miss the basics.
 
Doom112 said:
http://www.dsogaming.com/news/report-total-war-warhammer-runs-27-slower-dx12-nvidias-hardware/

This game shows that Nvidia is far better in DX11 in this game compare AMD in DX12.

http://www.pcgameshardware.de/Total...Specials/Direct-X-12-Benchmarks-Test-1200551/

CA never worked with NVIDIA on DX12 or took their input, which, one of the reasons the results are terrible on DX12. If this goes on then DX12 will have the same faith as DX10 ,therefore, developers need to take this brand sponsorship out and work with both parties.



http://wiki.totalwar.com/w/Optimisation_Blog


Once again sorry for my bad English.
Click to expand...

My 970 is also slower in Cryengine 5 running DX12 than it is DX11.

It seems to me they are just not native DX12 hardware, at least not in the same way AMD are.
 
Doom112 said:
http://www.dsogaming.com/news/report-total-war-warhammer-runs-27-slower-dx12-nvidias-hardware/

This game shows that Nvidia is far better in DX11 in this game compare AMD in DX12.

http://www.pcgameshardware.de/Total...Specials/Direct-X-12-Benchmarks-Test-1200551/

CA never worked with NVIDIA on DX12 or took their input, which, one of the reasons the results are terrible on DX12. If this goes on then DX12 will have the same faith as DX10 ,therefore, developers need to take this brand sponsorship out and work with both parties.



http://wiki.totalwar.com/w/Optimisation_Blog


Once again sorry for my bad English.
Click to expand...


Nvidia decided not to add Async compute to their GPU, because they do not want the tech to see a large adoption, instead they rely on pre-emption, that can mimic some of async functions, but not as efficient, still it is better on pascal than it was on maxwell, and it is more annoying for Devs to implement but again that's the small Nvidia deterrent for async, on top of the market share, because as long as it's complicated and the market is segmented, Devs would have more trouble comitting to it, which in return keep maxwell and earlier GPUs from Nvidia in the game, instead of losing 20-30% performance to the same priced AMD GPU.
and i dont think TW is done with their DX12 patch, they are still adding new features as they go, they dont have to add everything in one go
 
Last edited:
Even the Devs dont know about 1070/1080 DX12 support. Nvidia are being very quite on this it would seem.

Cn_Fi3x_PWYAANGCH.jpg
 
How long before Nvidia complain about this?


Manufacturer

GPU Async compute?
AMD RX480
Yes

AMD Radeon HD 7790
Yes

AMD Radeon HD 8770
Yes

AMD Radeon R7 260
Yes

AMD Radeon R7 260X
Yes

AMD Radeon R9 290
Yes

AMD Radeon R9 290X
Yes

AMD Radeon R9 295X2
Yes

AMD Radeon R7 360
Yes

AMD Radeon R9 380
Yes

AMD Radeon R9 380X
Yes

AMD Radeon R9 390
Yes

AMD Radeon R9 390X
Yes

AMD Radeon R9 Fury
Yes

AMD Radeon R9 Nano
Yes

AMD Radeon R9 Fury X
Yes

AMD Radeon Pro Duo
Yes

Nvidia GeForce GTX 950
No

Nvidia GeForce GTX 960
No

Nvidia GeForce GTX 970
No

Nvidia GeForce GTX 980
No

Nvidia GeForce GTX 980 Ti
No

Nvidia GeForce GTX Titan X
No

Nvidia GeForce GTX 1070
?

Nvidia GeForce GTX 1080
?
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom