Nvidia Pressuring Oxide developers to change DX12 Benchmark

[Inimigor]

Ok the situation is this:


Nvidia has a performance issue with the asynchronous compute, and then it's disabled, even though their driver claims support for it, and they want Oxide developers to change the Benchmark in order for it not to hurt Nvidia performance so badly.


What happened in this is that, Oxide has released a DX12 Benchmark where AMD gets a clear win, and Nvidia even has a worse performance when moving to DX12 according to the benchmark.

And Nvidia's PR has said that they don't consider this a proper benchmark since it has various issues, especially with MSAA.



Oxide's response to the situation is:

 

 

There is no war of words between us and Nvidia. Nvidia made some incorrect statements, and at this point they will not dispute our position if you ask their PR. That is, they are not disputing anything in our blog. I believe the initial confusion was because Nvidia PR was putting pressure on us to disable certain settings in the benchmark, when we refused, I think they took it a little too personally.

Oxide developer claims nvidia is placing A LOT of pressure for them to disable certains things in the benchmark in order to give Nvidia better performance. Bad for AMD. They also said that AMD wants to disable DX12 Asynchronous compute on Nvidia components.

 

Why that? you say.

 

Despite Nvidia's drivers claiming support, when Oxide's benchmark tries to use the method it hurts EVEN MORE the performance on Nvidia GPU's, so they disabled it on Nvidia side.

 

Also Oxide developers say that Nvidia has a bit more of a CPU overhead since their implementation of DX12 is tier 2, while AMD's is tier 3.

 

 

Personally, I think one could just as easily make the claim that we were biased toward Nvidia as the only ‘vendor’ specific code is for Nvidia where we had to shutdown async compute. By vendor specific, I mean a case where we look at the Vendor ID and make changes to our rendering path. Curiously, their driver reported this feature was functional but attempting to use it was an unmitigated disaster in terms of performance and conformance so we shut it down on their hardware. As far as I know, Maxwell doesn’t really have Async Compute so I don’t know why their driver was trying to expose that. The only other thing that is different between them is that Nvidia does fall into Tier 2 class binding hardware instead of Tier 3 like AMD which requires a little bit more CPU overhead in D3D12, but I don’t think it ended up being very significant. This isn’t a vendor specific path, as it’s responding to capabilities the driver reports.

 

Crazy huh?


AMD apparently won this so far, they have invested heavily in development and support for API's like Mantle and DX12, let's hope it pays off and we have some nice competition from now on!

------Personal Opinion------

While Nvidia could be unhappy with those performance issues I think trying to work together with the developers instead of AGAINST the developers would be so much better for their image as a company and even for solving the issue would be far less work.

 

The issue with the asynchronous compute seems to be withing a driver issue, since the GPU tries to use it, but won't run it properly, maybe bad code optimization (since it hasn't been used a lot) or something like that, all there is for us, is to wait and see what will happen!


Original article from Overclock 3D
http://www.overclock3d.net/articles/gpu_displays/oxide_developer_says_nvidia_was_pressuring_them_to_change_their_dx12_benchmark/1

[PlayStation 2]

*insert flame war here*

[Jerakl]

*insert flame war here*

 

/thread

[Majestic]

Oh boy, clickbait.

[Kinda Bottlenecked]

Why is so many shit happening? 

 

[That Norwegian Guy]

Nvidia is as gracious in defeat as a ballerina hippo at an ice rink.

[werto165]

"benchmark as a proper representation if DirectX 12 performance" if should be of. Why are they so bothered? I mean how many people are going to see this, most gamers don't care about benchmarks, sure it gives a good representation of the game but still.. 

[Aytex]

put som quotes bruh

[OAcesync]

What do Nvidia care so much about it instead of taking on board these problems with their drivers and trying to improve it because right now the looks of it AMD has a good advantage over Nvidia in DX12 like they promised

[colonel_mortis]

http://linustechtips.com/main/topic/11724-read-before-posting-in-this-section/

Unlocked to allow the OP to edit it.

[Icecube26]

DX12gate incoming

[Mister Snow]

Isn't this a repost?
 

http://linustechtips.com/main/topic/440575-oxide-responds-to-aots-conspiracies-maxwell-20-has-no-native-a-sync-compute/

[Dietrichw]

 

 

Also Oxide developers say that Nvidia has a bit more of a CPU overhead since their implementation of DX12 is tier 3, while AMD's is tier 2.

 

 

 

you got DX12 tiers mixed up. AMD is tier 3 and Nvidia is tier 2. 

 

Nvidia does fall into Tier 2 class binding hardware instead of Tier 3 like AMD

[PlayStation 2]

*insert flame war here*

oh wow i wasn't wrong

[6God]

Based on how I've seen other Nvidia/AMD threads on other parts go down, I'm sure this will be a civil thread with reasonable discourse from all sides and a mutual understanding that our preference in brand does not make us shills or fanboys and that we can hold mature discussion about the issue at hand without resorting to shenanigans. 

[Inimigor]

you got DX12 tiers mixed up. AMD is tier 3 and Nvidia is tier 2. 

Fixed

[Inimigor]

Isn't this a repost?

 

http://linustechtips.com/main/topic/440575-oxide-responds-to-aots-conspiracies-maxwell-20-has-no-native-a-sync-compute/

I think partly, the post is more about the issue with the API overhead, but the Nvidia pressure, apparently is something quite new.

[Curufinwe_wins]

No this is a complete repost of the other thread on this page written in one of the most inflammatory ways possible.

(And that's saying something.)

[Tigerbomb8]

I think partly, the post is more about the issue with the API overhead, but the Nvidia pressure, apparently is something quite new.

Nope, this is a re-post of http://linustechtips.com/main/topic/440575-oxide-responds-to-aots-conspiracies-maxwell-20-has-no-native-a-sync-compute/

[Prysin]

No this is a complete repost of the other thread on this page written in one of the most inflammatory ways possible.

(And that's saying something.)

how is this, or the other thread inflammatory?

 

It simply states a personal opinion, and sections of the article on OC.net

[colonel_mortis]

This topic (as well as the broader issue) is covered in the OP of this thread http://linustechtips.com/main/topic/440575-oxide-responds-to-aots-conspiracies-maxwell-20-has-no-native-a-sync-compute/