DirecxtX 12 for ArmA 3?

[calin_banc 19]

I hope you tested BF 4 in MP in some "hot spots" performance wise, 64 players on the same map and not in SP. SP is lite on the CPU and not representative. More so, BF 4 shows something nicely - a well written engine under dx11.1 is still significantly slower and choppier than a low level API.

In ArmA 3 it was just myself doing nothing in the editor. If in that scenario the game still "simulates" whatever, then Bohemia must be trolling. :D

Edited February 13, 2015 by calin_banc

[Sub-Human 10]

Idk guys, i think there should be an option to have Arma run either heavily CPU, or Heavily GPU. If they go straight GPU's, than i don't think my PC could handle it anymore. =/

I can guarantee you even a cheap graphics card will be able to process data faster than any $2000 Intel chip... ArmA should focus more on GPU, perhaps even AI calculations and physics engine.

[brightcandle 114]

I can guarantee you even a cheap graphics card will be able to process data faster than any $2000 Intel chip... ArmA should focus more on GPU, perhaps even AI calculations and physics engine.

Just be aware that the time per frame for AI and physX is pretty minimal, less than 1ms. AGM currently think the 3ms cost they add to each frame is nothing so you aren't going to find many people care about 5% of the frame time at 60 fps. Most of the time is not spent in these periphery activities in a multiplayer game.

I know its hard to read the information I have provided, its designed for developers and this being the first profiler output you may have seen its tricky to understand it, so in the absence of understanding you'll have to trust my word for it when I say this isn't the issue.

[calin_banc 19]

Some more talk about DX12 -

Edited February 14, 2015 by calin_banc

[R3vo 2654]

Some more talk about DX12 -

That was very informative. Thanks for that.

[DancZer 65]

Some more talk about DX12 -

Omg! Omg! Thank you!

I think that, if BI wouldn't make the Arma 3 with full DX12 support, they would have a competitor soon.

[calin_banc 19]

A direct competitor is not likely to appear, there isn't that much money into this market to worth the risks, not at the moment. However, from a technical stand point, sure; more and more engines will get close. Right now RSI is working on heavily modifying the Cry Engine for their game and at this point it looks like it's going far beyond what's ArmA doing.

Back to DX 12 - http://www.dsogaming.com/news/dx12-versus-dx11-over-100fps-difference-on-unreleased-gpu-in-new-test-way-beyond-console-stuff/

Did a test of DirectX 11 vs. DirectX 12 on an unreleased GPU with an 8core CPU. DX11: 13fps, DX12: 120fps. Lighting and lens effects.

[Misfit Leader 1]

Sorry for replying an old thread but now that we know that DX12 will be Windows 10 and that every Windows 7 (at least for private use) should be able to upgrade to 10 free of charge...

Do BIS think of adding DX12 support ? Maybe not all the new stuff but...

The most interesting thing in it is the support of the "mantle like" optimization to lower the CPU load. And that should be the best addition to arma which is very CPU "bottlenecked", so freeing a little bit the CPU could help the game ?

Woops : Haven't seen a new thread opened for it

-> http://forums.bistudio.com/showthread.php?188022-BIS-CEO-Marek-Spanel-Wants-DX12-for-Arma-3!

Maybe moderation should close this one, and pinpoint the new one in the first post :rolleyes:

Edited February 19, 2015 by Paul-Hewson
new thread spotted

[Maio 293]

Threads merged.

[DancZer 65]

DX12 is going to have no impact on the game.

I don't agree with you. The game engine(data processing, draw preparation) is designed to use DX11 with consideration of the API limitation. If they move to DX12 probably they will redesign the flow of the engine, which you can't measure.

[chortles 263]

Right now RSI is working on heavily modifying the Cry Engine for their game and at this point it looks like it's going far beyond what's ArmA doing.

I hope that you meant a new company and not RealTime Immersive, because otherwise I can find that one post where a BISim support guy practically crowed at how RTI didn't work out as a "VBS killer"... albeit for reasons that have jack-all to do with Arma.

[calin_banc 19]

Robert Space Industries -> Squadron 42/Star Citizen.

[chompster 29]

A direct competitor is not likely to appear, there isn't that much money into this market to worth the risks, not at the moment.

http://joinsquad.com/

hopefully these guys are.

[Jackal326 1181]

http://joinsquad.com/

hopefully these guys are.

Handy if you want PvP, but as they don't plan to do any co-op (as far as I could see from their site), I'll stick to ArmA.

[bad benson 1733]

http://joinsquad.com/

hopefully these guys are.

that's project reality basically. so one PvP mode (AAS) instead of the whole of arma. not saying UE4 is not capable of it. just that squad is, afaik, based around realisitc PvP only. looking forward to it though.

i don't see what the big deal about "what arma does" is supposed to be anyways. arma doesn't do shit. it's modders and scripters that do the interesting stuff. gamemodes and features that go beyond what is basically a more clunky than average and bigger than average shooter with vehicles.

i don't see anything mind blowing happening while it actually performs at a non-painful level too. having good FPS on an empty map doesn't create good gameplay. people like to juggle these huge numbers when it comes to arma but, if you would really analyse it and see under which circumstances it actually performs at an "acceptable" (by normal standards like comparing to other games...) rate, it would be a quite sobering experience.

the only thing that arma as a game does different in a bigger way is AI. and if you played against them since ofp, they won't blow your mind. the opposite actually. not to mention how they drive (or don't). they fly ok. but there are no obstacles so it's hard to judge. they still suck at landing. their limited behaviors are understandable since they have to scale up and keep doing stuff. but after all these years being basically the same at the core, it's really nothing to brag about (only enterable buildings on the map but AI ignores them...).

the one thing that gives arma long term "playability" is the editor and its moddability. it makes it basically like an engine with a lot of starter content (relatively :p) for hobbyist to play with. the great side effect for the devs is that all the content those hobbyist constantly push out increases the value of their basically updated ofp by a million and makes it more than it is just as a game. even without the much appreciated progress the current devs are making, it would be a constant construction site with the constant promise of more stuff to discover or general improvement. early access at its best.

adding that mission editor was the single most lucrative decision the original devs ever made. in my eyes the mission makers basically make the game survive. i mean the reason mods are so limited and these days mostly focussed on fixing the game are mainly the tools, documentation and arma overall being not very moddable (story of my life :D) so props to anyone keeping at making mods for arma. but i sometimes feel that the mission makers are taken for granted because they always did this job. but now more than ever arma would literally be dead without the community designing and creating its MP.

looking at how UE4 is going for large scale support and just generally speaking: more modding allowed on other engines = death of arma. i really think that the military aspect has far less to do with arma's success than it being a sandbox. realism is an intriguing thing but arma isn't even that realistic. what makes it good on the gameplay side is it being hardcore. challenge = fun. streamlined games simply get old much faster. let's hope the industry relearns that some day :D

anyways. too much beer = too many words.

[windies 11]

Pretty good points Benson.

[chortles 263]

adding that mission editor was the single most lucrative decision the original devs ever made.

As something of a tangent, there's something tragicomic about how the current creative director almost immediately downplayed the possibility of a 3D editor arising from Zeus, only for BI to recently announce that work was going ahead on it, if only because I'm left wondering whether that meant that BI actually took that long to get a 3D editor to "prototyping" (read: having starting much earlier than reported) or whether instead it took that long before they had the confidence in its viability to announce its existence...

let's hope the industry relearns that some day :D

I'm sure that industry professionals are perfectly aware, but the sea change would be if suits come around to thinking that "challenge" can make them more money now (not "in the long run") than "not challenge"... after all, some people actually do show up for the interactive movie (I'm looking at you, Assassin's Creed and Dynasty Warriors fandoms), while the trend has cultivated an environment where Dark Souls could be sold literally on the basis of "prepare to die". :lol:

[megagoth1702 252]

Arma's performance won't change with any DX version. It's an old engine, designed to run on one thread mainly. Nothing is going to change that, not even DX25. BIS will need a lot of time and ressources to code a new engine based on multithreading (like cryengine, frostbyte) but that's not happening any time soon.

[DancZer 65]

Arma's performance won't change with any DX version. It's an old engine, designed to run on one thread mainly. Nothing is going to change that, not even DX25. BIS will need a lot of time and ressources to code a new engine based on multithreading (like cryengine, frostbyte) but that's not happening any time soon.

Are you an Ex BI developer or what? :D

[St. Jimmy 272]

Are you an Ex BI developer or what? :D

Well world rendering has never been the biggest issue in Arma but the scripts and AI. It can ease things up but the same big performance drop will happen when AI, scripting and stuff are added. RV needs more than just a new DX. BrighCandle said it very well few pages back and in some topic. Arma 3 isn't really even hitting the limit of DX11 according to those benchmarks.

[calin_banc 19]

Go on a empty map and crank up de view distance, object view distance and object quality. See frame rate and gpu usage drop. It's pretty much text book limitation.

[I give up 152]

Well world rendering has never been the biggest issue in Arma but the scripts and AI. It can ease things up but the same big performance drop will happen when AI, scripting and stuff are added. RV needs more than just a new DX. BrighCandle said it very well few pages back and in some topic. Arma 3 isn't really even hitting the limit of DX11 according to those benchmarks.

Rendering is the biggest issue with Arma 3.

Mainly due to the fact that it uses DX11 in a engine that has limitations in matters of multithreaded/core usage.

Only those who dont have a clue about DX11 architecture, demandings and needs may think that is possible to have it running (within acceptable conditions) on a basically single threaded environment.

[bad benson 1733]

Rendering is the biggest issue with Arma 3.

while i'm not an expert by any stretch on the subject, i kinda doubt that. seems more like it's the other way around. or less simplistic, how sim and render are too tied together.

[fareast 20]

even if performance gain is minimal..with dx12 we can have more light sources and maybe 12k view distance without performance hit...try flying with heli above kavala at night..light popping everywhere because of dx11 limitation..also PAPI,REIL,ALS and plane/heli navigation light can have real light source now instead of the fake blinking whatever they have right now..last but not least PIP scopes might be possible now

[brightcandle 114]

Right now Arma 3 is using about 2000 draw calls even with high rendering distance. As you increase the render distance you'll see the profiler show the rendr section of the profile increase, however when you look in the GPUView output you wont see a corresponding increase in DirectX API overhead. Thus I came to the conclusion some months ago that the game is rendr (and sim) limited but its not DirectX's fault, BI are no where near the limits of the API. Infact what we see is that the rendr section of the game is the only part that shows any amount of multithreading at all, but most of the time is in BI code and not in DX 11. I have done everything I can think of to determine the root cause of the problem (as a developer with years of experience in profiling and fixing problem programs incidentally) and my end conclusion is that all of the problem lies in BI's code. The primary factors for poor performance on the client is the games simulation, which includes running the scripts but not the AI (which is mostly server side) and the rendering process but this is mostly in BI code and not in DX11. I have all but categorically proved it with profiling output, Microsoft debugger tools and draw call estimations all of which tells us a lot about what Arma is doing when its running slow. The next step is BI needs to fix the game, and DX12 isn't going to do a lot in this case, it will improve the frame rate because it will reduce an overhead in the main thread and it might give BI some more avenues for parallelism, but fundamentally the problem is 100% with what BI has done and not the API they are using.