DCS A-10C Screen Shots - 16 Aug 2010

[aaron886]

It's more just a good sign that they're keeping up with the "latest." All they need is true utilization of multiple cores and we're in 2010!

[kingneptune117]

It's more just a good sign that they're keeping up with the "latest." All they need is true utilization of multiple cores and we're in 2010!

 

Is it possible to release this kind of functionality through a patch?

[636_Castle]

Is it possible to release this kind of functionality through a patch?

 

I don't see why not. You can manually utilize all CPU cores by alt-tabbing out and resetting the affinity in the task manager. There's also 3rd party launchers at lockonfiles.com that do this as soon as you launch the sim. It's not 100% optimized, though.

 

As for DX11, I'm thrilled about it. Finally an excuse to ditch my DX10 card that I'm so fed up with! I've had this 9800GTX+ since I built this rig, and it bottlenecks my CPU more than an espresso straw in a milkshake. Plus support for tessellation. I think many current DX9 users are going to be a little frustrated though.

Edited August 19, 2010 by 636_Castle

[EtherealN]

I don't see why not. You can manually utilize all CPU cores by alt-tabbing out and resetting the affinity in the task manager. There's also 3rd party launchers at lockonfiles.com that do this as soon as you launch the sim. It's not 100% optimized, though.

 

The 3rd party launcher you are talking about serves very little purpose, since the products now have affinity to all cores even without it. This was only a factor during DCS:BS 1.0.0. The patch to 1.0.1 sorted this.

 

However, note that this does not mean you "utilize all CPU cores" in the sense that they are all used at the same time. You do not make the process thread split into multiples. All you do is you allow the OS to switch between cores, something that can be done really fast to ensure that it works in the most free core.

 

Also, this is not a question of "optimization". It's a question of re-implementing large parts of the code. This is being worked on though - for example the new sound engine was implemented as a separate thread, and I would suspect that such pieces of the application's intestines will be continually updated and placed in their own threads in that manner as and when they are implemented.

[636_Castle]

I was stating the launcher tool isn't optimized to give you full functionality. Sooo.. :noexpression:

[Callsign.Vega]

Flight Sim's being heavily CPU Dependant, really need to go to multi-threaded to utilize 4-6 cores. Sim's like Black Shark really only use 1 core for the main simulator thread (no sound etc), really hampering FPS. My GPU barely reaches 50% utilization at 2560x1600, and only 1 of 4 cores is maxed out while playing BS and I get a measly ~25 FPS. Game designers really have to step it up and fully utilize current hardware.

[KungFuCharlie]

Parallel processing isn't a simple "mod" to a project. A game engine would have to be designed from the ground up with parallel processing in mind. Creating a multi-threaded application is a little tricky but easily do-able. Creating an application the truly supports parallel processing is a whole new challenge.

 

I really wish game / sim designers would get on the parallel processing bandwagon soon but the cost (engineering wise) is probably out of range for many. Lets just hope ED is getting enough $ in military contracts to make this happen!! :D

[jctrnacty]

Flight Sim's being heavily CPU Dependant, really need to go to multi-threaded to utilize 4-6 cores. Sim's like Black Shark really only use 1 core for the main simulator thread (no sound etc), really hampering FPS. My GPU barely reaches 50% utilization at 2560x1600, and only 1 of 4 cores is maxed out while playing BS and I get a measly ~25 FPS. Game designers really have to step it up and fully utilize current hardware.

 

I have a deep faith in ED in multicore support, because if someone is able to program something so sophisticated and complicated like BS then multicore support is just a question of time. The problem is that very often you have to dump most of the work and begin from scratch. Because you need to build whole program with MC support in mind from the begining. It is not easy to add it later.

[Sticky]

I have a deep faith in ED in multicore support, because if someone is able to program something so sophisticated and complicated like BS then multicore support is just a question of time. The problem is that very often you have to dump most of the work and begin from scratch. Because you need to build whole program with MC support in mind from the begining. It is not easy to add it later.

 

How many years would that take? Wouldnt you much rather have a new DCS plane each year that would take? Who knows how many years it would take, three, four years? Would you exchange those three or four DCS planes for the one new engine? How much faster would that engine be anyway? Would it be worth it?

 

Maybe if they have a seperate group working with the new engine while the larger group keeps doing the DCS modules, but does ED have these recources? I hope we will see true multicore support one day, but I dont see how that can be done without sacrificing upcoming releases.

[Sticky]

If the choice is having four DCS modules in four years BUT having to buy a new faster cpu two times during these four years VS nothing at all for four years what would you choose? Ofcourse four years is just a wild guess, it could be anything.

[Boberro]

I dont see how that can be done without sacrificing upcoming releases.

 

Neither do I. Anyway new engine is more important.

[Sticky]

What if you had a lethal disease and only had a couple of years left to live?:cry:

[jctrnacty]

How many years would that take? Wouldnt you much rather have a new DCS plane each year that would take? Who knows how many years it would take, three, four years? Would you exchange those three or four DCS planes for the one new engine? How much faster would that engine be anyway? Would it be worth it?

 

Maybe if they have a seperate group working with the new engine while the larger group keeps doing the DCS modules, but does ED have these recources? I hope we will see true multicore support one day, but I dont see how that can be done without sacrificing upcoming releases.

 

1. I dont know

2. Yes I would

3. Maybe 3 or 4 or 5

4. No I wouldnt

 

If I say i have a deep faith in ED and MC support, then it doesnt mean I want it right here and right now. I think you misunderstood something. So there is no need to be offensive immediately.

[Sticky]

I wasnt trying to be offensive, I have no clue why you would think that, maybe I need to rethink how I express myself? Anyway you wrote that its a question of time, and I just wanted to know how much time you would think was an acceptable sacrifice, for discussion purposes, nothing else LOL

[Sticky]

Also: Time = Money ;)

[nscode]

Edited August 20, 2010 by nscode

[Speed]

It's more just a good sign that they're keeping up with the "latest." All they need is true utilization of multiple cores and we're in 2010!

 

Well, according to that link Poko posted, http://blogs.amd.com/play/2009/09/09/directx-11-–-what-to-expect/, then SOME multicore support is part of what directx11 provides.

[sinelnic]

How many years would that take? Wouldnt you much rather have a new DCS plane each year that would take? Who knows how many years it would take, three, four years? Would you exchange those three or four DCS planes for the one new engine? How much faster would that engine be anyway? Would it be worth it?

 

Maybe if they have a seperate group working with the new engine while the larger group keeps doing the DCS modules, but does ED have these recources? I hope we will see true multicore support one day, but I dont see how that can be done without sacrificing upcoming releases.

 

Excellent points. From my experience the key part would be if they´d have to start entirely from scratch, making the new engine project basically a 3-yr R&D project + 1 yr for adapting the content to the engine and bug fixing, or if they have already done experiments, prototyping, and/or have gone through the low level design exercise that is required.

 

If they did do the R&D part during the last four years (say since the establishment of the Core technology in the consumer market), I´d wildly guess a 2 year work.

 

I would think that going multicore is not the only benefit of a new engine. DCS could improve fidelity immensely by actually modelling the atmosphere, as a dynamic fluid rather than the current static simplification, allowing for thermal currents, wind deflection in mountains, aircraft induced turbulence, all things that would take the overall flight "model" to the next level on par with high-end systems. Besides that, dynamic modeling of air temperature and humidity would allow for a very realistic weather system.

 

Couple that with the enormous room for improvement in the visual simulation given by the new technologies, with particle-based dynamic clouds, smoke and precipitations; using DX11 tesselator for terrain (maybe going procedural) that could allow, for instance, to have very detailed cities where all non-building objects become 3D instead of textures at close range (i.e. cars, traffic lights, small plants, crops, etc) while at the same time, the long-distance textures are pre-rendered from those objects (so forests don´t pop up from nowhere); using newer techniques for object drawing (i.e. full DX11 geometry instancing); collidable trees (why not); and finally, dynamic terrain lighting with mountain, building and vegetation-casted shadows on all objects plus dynamic higlights in the terrain according to the sun position (so that sun-bathed forests look like a sun-bathed broccoli if you get the picture).

 

So I would think a beefy engine update ED-style, would be quite something, and something I would personally pay for and enjoy as much as a new plane. The level of detail in the plane simulation, at least for me, means that I need to get close to professional standards to fly the machines well, with a true sense of satisfaction, and that takes years, not days, to obtain, so I´m in no hurry for new planes if I´m getting something like I described above instead.

 

Couple that with a more complex and realistic mission evaluation system, career mode, dynamic campaign... hell I´d pay a lot.

[Surf]

Nice Wags. Adding DX11 haze/smoke/water capablities will be great,

and improve an already compelling enviroment.

 

-Surf

Edited August 20, 2010 by Surf

[Python]

How many years would that take? Wouldnt you much rather have a new DCS plane each year that would take? Who knows how many years it would take, three, four years? Would you exchange those three or four DCS planes for the one new engine? How much faster would that engine be anyway? Would it be worth it?

 

Maybe if they have a seperate group working with the new engine while the larger group keeps doing the DCS modules, but does ED have these recources? I hope we will see true multicore support one day, but I dont see how that can be done without sacrificing upcoming releases.

 

 

I know i for one would be a much happier chap getting a new DCS every year, it means we may get round to a classic British attack aircraft while im still young, variety is what we all want ofcourse. I think alot depends on how DCSWH performs over the range of our systems, im hoping for the best, i suppose we will fing out soon enough.

 

 

If I say i have a deep faith in ED and MC support, then it doesnt mean I want it right here and right now. I think you misunderstood something. So there is no need to be offensive immediately.

 

I didnt see anything offensive in sticky's post?? were all friends here :)

[Martillo1]

What if you had a lethal disease and only had a couple of years left to live?:cry:

 

Then you are screwed :cry:

[TheMoose]

DX11 :wub: Thanks ED

[lce]

I know you guys are busy, but maybe it's a good idea to update the DCS front page/ news section with some of this info!

http://www.digitalcombatsimulator.com/ looks kinda old not having recent news/screnshots. :)

[Tangolima600]

[metalnwood]

Bloody hell is there a person under that. Reminds me of hawk from buck rogers...