Performance in 2.5 Cacasus map

Can somebody of the testers (who has experience with both version) tell us, how the performance with all the new graphical features on the caucasus map is affected in comparison to 1.5.8.

I would guess that it's a big hit in performance.

I would guess that it's a big hit in performance.

Well you'd be surprised...

I'm not a tester myself, but there's a good performance difference between the current 2.x version and 1.5.x that bodes well for 2.5

Besides, the new Caucasus map will have speedtrees which are very fast indeed and that alone should make for a very positive performance differential between 1.5 and 2.5

Of course YMMV (your mileage may vary).:smartass:

Keeping in view the videos they shared I think it will require 1080ti or 1070 sli to run on high graphic settings. 1070 sli manages well in 2.2 on High settings. should be able to handle 2.5 also. I was able to make sli work in 2.2 and hopefully 2.5 should also work.

I hope it will be better than 2.2 Normandy.

I hope it will be better than 2.2 Normandy.

I am with you there.

[...]

Besides, the new Caucasus map will have speedtrees which are very fast indeed and that alone should make for a very positive performance differential between 1.5 and 2.5[...]

could someone quickly explain why speedtrees are supposed to be so light on performance? i think the SPEEDtree name comes from the relatively low amount of time needed to generate unique trees and populate levels with them.

when rendered they are still nothing less than some polygons and textures, while the old caucasus trees are just cardboards/sprites.

i can't see how a "speedtree" could render faster than a sprite...

could someone quickly explain why speedtrees are supposed to be so light on performance? i think the SPEEDtree name comes from the relatively low amount of time needed to generate unique trees and populate levels with them.

when rendered they are still nothing less than some polygons and textures, while the old caucasus trees are just cardboards/sprites.

i can't see how a "speedtree" could render faster than a sprite...

Maybe have a look at their website. Seems they've put some work into it.

https://store.speedtree.com/

Can somebody of the testers (who has experience with both version) tell us, how the performance with all the new graphical features on the caucasus map is affected in comparison to 1.5.8.

I would guess that it's a big hit in performance.

None of the testers will answer this question due to N.D.A.,

However You can search for Videos and Posts by Wags and get an idea of Performance.

Performance is slightly better then we have now. Matt has said that in one of his vids i remember. Once vulcan API is integrated there is a near 40 procent fps boost. This is what i know. :D This is the reason i wait too buy a new Gpu.

I know it won't happen overnight but I'm happy to hear that ED plan to take the Vulcan route instead of DX12!

Who knows, maybe we will see Mac/Linux versions of DCS down the road ;)

I know it won't happen overnight but I'm happy to hear that ED plan to take the Vulcan route instead of DX12!

 

Who knows, maybe we will see Mac/Linux versions of DCS down the road ;)

Bloody hell this hype is bringing a lot of old timers back, nice to see you mate :)

I'm happy to hear that ED plan to take the Vulcan route instead of DX12!

So am I: OpenSource > proprietary, and my best friend was a part of the Khronos group :-)

Any info on Vulcan ETA?

Any info on Vulcan ETA?

2.5 will be here within 18 days I don't have any idea if it's supposed to be Vulcan based.

But it is promised before the end of this month!

The only thing I've read about Vulcan in these threads is all still speculation.

I looked all over the place and all I've seen is discussion.

2.5 will be here within 18 days I don't have any idea if it's supposed to be Vulcan based.

But it is promised before the end of this month!

The only thing I've read about Vulcan in these threads is all still speculation.

I looked all over the place and all I've seen is discussion.

No, 2.5 is still the good old DX11 AFAIK. :D

No, 2.5 is still the good old DX11 AFAIK. :D

Not too long anymore (hopefully) :D

Bloody hell this hype is bringing a lot of old timers back, nice to see you mate :)

Thanks Cibit.

Regarding Vulcan ETA I don't expect it anytime soon, heck 2.5 that is the DX9 > DX11 transition isn't out yet.

But still happy ED will avoid Microsofts DX12 lock-in chains. MS have been doing a lot of good stuff in the past (been using Windows since 3.0) but they are heading in the wrong direction with W10 - IMO of course.

Personally I will stay with W7 until I can't run it anymore and after that it will be Linux all the way unless we see MS going back to their roots with "Windows Whatever".

Highly doubt it tho since MS soul departed with Bill Gates many years ago. :prop:

I hope it will be better than 2.2 Normandy.

I hope it'll be at least as good as v2.2 NTTR since it runs absolutely great on my oldish rig ;)

From what I remember, Sith Spawn, on couple of occasions, in the map comparison threads, has said that he is seeing a better performance than in the current version

Good news, then. Wowsa :thumbup:

I read on here somewhere by someone that Vulcan API is due by the end of 2018. Don’t think it was official though.

It is impossible to say how well does something run from the YouTube video. You would really need at least some benchmarking like playing back the same mission type (about same position of waypoints etc) and then see the FPS and RAM consumption at different resolutions like Full HD and 4K so one can do estimations.

But in the F/A-18C video Wags said he had all maxed out and specs were 1080Ti and IIRC a i7-77000k with 32GB RAM. And he said that he had not a single hiccup (the Live Stream had couple ones).

But such a system is anyways top notch already.

And IMHO it is just wrong that we are expecting that everyone should have a 1080Ti or even 10xx series, as today things should be tested against a 8-16GB RAM running with a 4-core CPU (few year old) with a GTX9xxx series.

Throw there a VR setup and it becomes a trouble as even with 4K you can get great flight experience as you only need that 30+ (very) stable FPS to get all smooth, but with VR you need to get at least 45 FPS.

I can run a Normandy in 4K all maxed out for about 110-140 FPS by average, but once I hit the VR and the default VR setup in settings, I crawl in 12-15 FPS.

Even in a NTTR it is about same, VR just eats every single bit but 4K runs perfectly smooth graphically as long you don't have AI units operating or FPS drops seriously below 20 FPS when just 20 units are on ground.

In the current 1.5.8x in all maxed out and 4K I can get 180-240 FPS but throw again couple ground units and we are in slideshow. Take a VR and things are sub 30 FPS.

I have already built personally myself a two different setups, current one being i7-8700K and decided to try a cheapest Threadripper (1900x) but turned by head as what use it is when DCS is just using a single thread for all other than audio, and the Vulkan is required to get the other CPU cores fully utilized so GPU doesn't get throttled by some ground units interfering the graphics performance.

This is one reason why I as well prefer a 4K as it is easier to get excellent graphics rendering (solo) compared to VR.

But I am now waiting the 2.5 Early Access release so I can check what my current setup can do, even I don't really expect much different in VR, as I can always revert to 4K for better performance if so comes.

I read on here somewhere by someone that Vulcan API is due by the end of 2018. Don’t think it was official though.

In official 2017 December newsletter it is written as:

We will also be working on improvements to clouds, explosions with improved proximity damage, virtual reality, spotting system, network play, Forward Looking Infrared (FLIR), air-to-air missiles, and performance optimization with the inclusion of the Vulkan API. In parallel, we will continue to support and improve our existing modules.

With the pending release of DCS World 2.5 which will unify our projects and several exiting new modules, we see 2018 as being a great year to be a DCS World fan!

https://forums.eagle.ru/showpost.php?p=3338291&postcount=124

So keyword might be "inclusion" as it doesn't promise big improvements etc. We don't know how much Vulkan can help in performance as it depends how DCS is coded. But at best case scenario we should see 40-50% performance increasement. So VR to go from 22-24FPS to 30 FPS would be a huge thing.

First off, I'm not running a VR set of any kind here, just a single flat screen. But still...

I wonder what exactly is it that slows some setups to a crawl with just a couple of active units, since they cause no hiccups whatsoever for me - and I'm running, believe it or not, an ancient i7-2600K / GTX980 / Win7 rig here. Both the CPU and the card are overclocked so that of course helps, and I'm using a modded config file set with reduced draw distances too; but still, you'd expect a rig this old not being able to run DCS at all when even the beefiest modern computers seem to struggle. And yet, I'm seeing a constant vsynced 60FPS almost everywhere regardless of how many ground units I've got active!

I'm beginning to think this may be because of how I've configured the OS. Mine has 25 services started at boottime all told and no other background processes whatsoever so it is a slimmish setup indeed, plus all the drivers, DLLs and whatnot are up to date too. So how about looking into that first if you're experiencing problems?

most of you guys have monster machines, I don't know why are you so worried about it. It is bound to be performing better as it is newer and more optimized.

people like me, with GTX560Ti, should be asking such questions. Although, I am not worried. Even if it is not running better for me, I have an old card, so should consider upgrading