Question about Vulkan Expectations

[Dangerzone]

I know that ED are busy porting DCS from DirectX over to Vulkan.

I was just wondering what are the main benefits that we can hope to get from Vulkan once this has been done?

In particular, can we expect to see noticeable improvements with the CPU usage/binding, or is Vulkan only really for GPU benefits only? 

I know that a lot of this will be speculation - however given I have zero understanding of DirectX vs Vulkan - just curios as to what are some of the more likely benefits we can hope to see from anyone who's saavy in understanding that area.

[Mr_sukebe]

The last comments I read from ED was to assume no benefits initially.  Probably a very prudent approach.

My understanding is that CPU loading is typically lower for apps using Vulkan when compared to go DirectX.

Additionally, when Xplane11 added Vulkan support, I saw a big improvement is smoothness and frame rate.

In short, long term, should be a good thing, but play safe with expectations when it first goes live.  With the complexity of DCS, chances are that there will be a few snags.

 

[Raven (Elysian Angel)]

I don't know if you have seen following video, but I think it sums up pretty well what we can expect:
 

Personally, I don't expect any benefit with the first iteration either. It will likely be purely to test compatibility and to make sure it doesn't break anything or causes unforeseen issues.
Later on ED will refine it, and as Wags has stated in multiple interviews: Vulkan will allow them to add a lot more effects to DCS which are currently not possible - including raytracing but also FSR 2/3/4 for example (DCS currently only supporting FSR 1 is due to the limitations of Direct3D 11).

Additionally, you can check out this thread here where people have been discussing the same questions you have for years:
 

 

[Dangerzone]

19 hours ago, Mr_sukebe said:

Additionally, when Xplane11 added Vulkan support, I saw a big improvement is smoothness and frame rate.

This for me is encouraging. My 2 issues with performance doesn’t have anything to do with getting better graphics. One is a smoother consistency (get away from all these things that cause jitters, stutters, etc) and the other would be the rapid degrading in performance with the more unit count on the map. 

If Vulcan doesn’t give us any more fps, but we do benefit from being stable, and being able to do more (not graphics per se, but objects) - that is a huge win. 

Thanks! 

[SharpeXB]

On 6/21/2025 at 8:48 AM, Raven (Elysian Angel) said:

I don't know if you have seen following video, but I think it sums up pretty well what we can expect:

That video was made three years go before DCS adopted multi-threading. So the Vulkan advantages for single CPU systems don’t seem as relevant anymore.

[Raven (Elysian Angel)]

59 minutes ago, SharpeXB said:

That video was made three years go before DCS adopted multi-threading. So the Vulkan advantages for single CPU systems don’t seem as relevant anymore.

Sure, but it's been well established by ED that you can't have Vulkan (in DCS) without multi-threading so the one who made that video likely knew that? Also, DCS still has a main/synchronisation thread that should benefit from single-thread optimisations I'd think (or at least the reduced overheard), but perhaps I'm talking out of my behind  

Edited June 22 by Raven (Elysian Angel)

[Dragon1-1]

On 6/21/2025 at 1:26 PM, Mr_sukebe said:

Additionally, when Xplane11 added Vulkan support, I saw a big improvement is smoothness and frame rate.

This was because before Vulkan, the only option for Xplane 11 was OpenGL. So obviously you'd see a major benefit in this case, because they moved to an API that isn't ancient.

DCS will likely not see anything nearly that dramatic, however, due to the way VR is implemented in DX11, we'll possibly see a performance increase, possibly a major one, for VR users specifically.

[Mr_sukebe]

On 6/21/2025 at 12:26 PM, Mr_sukebe said:

The last comments I read from ED was to assume no benefits initially.  Probably a very prudent approach.

My understanding is that CPU loading is typically lower for apps using Vulkan when compared to go DirectX.

Additionally, when Xplane11 added Vulkan support, I saw a big improvement is smoothness and frame rate.

In short, long term, should be a good thing, but play safe with expectations when it first goes live.  With the complexity of DCS, chances are that there will be a few snags.

 

 

well spotted!

I’ll get my coat…

[av8orDave]

Has there been some news that I've missed regarding Vulkan implementation? It has been discussed for probably five years or more without any actual evidence of the work, so I'm skeptical that we'll ever see it. I hate to be so negative and I'd be glad to be proven wrong.

[BIGNEWY]

15 minutes ago, av8orDave said:

Has there been some news that I've missed regarding Vulkan implementation? It has been discussed for probably five years or more without any actual evidence of the work, so I'm skeptical that we'll ever see it. I hate to be so negative and I'd be glad to be proven wrong.

Its very difficult to show code as news and make it interesting. We will share more news when we are ready to, but for now work continues and is a high priority for us. You can choose to believe it or not, we can only tell you that our plans have not changed and vulkan implementation is something we are working towards. 

thank you  

[av8orDave]

Thanks for the reply, and I do trust and believe the team is working on it. The reason for my negativity is that this process has been what I would say is really, really long. The link below is from 2021, and estimated implementation of Q3 of that year.

 

Edited June 23 by av8orDave

[Dangerzone]

1 hour ago, av8orDave said:

Thanks for the reply, and I do trust and believe the team is working on it. The reason for my negativity is that this process has been what I would say is really, really long. The link below is from 2021, and estimated implementation of Q3 of that year.

Take note of the small things. BN said that they're working "towards" it. It sounds like it's less down the progress path than I first anticipated. I recall when DLSS was being worked "on" and they were actively testing it...  it featured in one of the "20xx and beyond" video's as small text down the bottom. Then we knew that it was "soon".

Dynamic Campaign engine has been worked towards (and is now being worked on) for... 10 years IIRC? So I understand you're negativity - however I would see it more as just been realistic than negative. DCS is a huge and complicated program, and these tasks are not something that are measured in days, weeks, months, or even in some cases years.    The best way to avoid disappointment is to understand that these are things that will come eventually - but yes - we could be talking in measurements of half decades for progress, so just alter your expectations accordingly. 

I would expect that when it's within a year or so of release, we'll see a video by DCS with a small text down the bottom "Filmed using Vulkan" or similar. Untill I see that - I don't hold my breath. 

[PawlaczGMD]

On 6/21/2025 at 9:48 AM, Raven (Elysian Angel) said:

I don't know if you have seen following video, but I think it sums up pretty well what we can expect:
 

Personally, I don't expect any benefit with the first iteration either. It will likely be purely to test compatibility and to make sure it doesn't break anything or causes unforeseen issues.
Later on ED will refine it, and as Wags has stated in multiple interviews: Vulkan will allow them to add a lot more effects to DCS which are currently not possible - including raytracing but also FSR 2/3/4 for example (DCS currently only supporting FSR 1 is due to the limitations of Direct3D 11).

Additionally, you can check out this thread here where people have been discussing the same questions you have for years:
 

 

FYI, you can use OptiScaler to redirect DLSS inputs to FSR3. It works. Therefore, I'm not sure it's a technical limitation.  FSR4 requires a 9000 series GPU, but I understand it would also work. It seems like ED could implement FSR3 and 4 now if they wanted to.

[j9murphy]

My limited understanding of converting to Vulcan would suggest that it will (initially)...

• bring more consistency to framerates (not necessarily more fps, but smoother fps
• create the basis for making further performance optimizations (this is a long road and one that only STARTS with implementing vulcan.) 
• allow ED to make other enhancements without negatively impacting frame rate as much 

They could also choose to take advantage of some of these things before they release it to us, meaning that they take a little longer but we get more tangible benefits when its released.  - most of this was said above:)

Edited June 25 by j9murphy