DCS World 2.0 and New Maps Discussion Thread

[Abburo]

Maybe you know this "documentary" ... or not...

Anyway, real Nevada Operation Red Flag :

 

http://youtu.be/-7zk_2S5esE

[shagrat]

With the emergence of more substantial WWII content in DCS, they might become interested. A recent article on Rise of Flight: Channel Battles Edition was actually celebrating it as being a detailed and unforgiving, but also accessible simulation:

 

Review

 

After Mick Schnelle in his role of professional "Ludo" (Mensch-ärgere-dich-nicht) Player decided that "Chess" is much to complicated and strategic to master for average players, I guess they are a bit biased!?

As he wasn't even able to start the A-10 without prior reading the 400 pages manual - I would have used WIN + HOME, but, yeah I'm a chess player, you know - so I wonder if anything that needs more than two or three switches to start or operate will get their laudatio...

 

Gamestar on DCS: A-10C Warthog

http://www.gamestar.de/spiele/digital-combat-simulator-a-10c-warthog/test/dcs_a_10c_warthog,46535,2322050.html

[upupandaway]

Sent you a PM. It´s getting a little too off-topic I think ;)

[X93355]

Edge and Upgrades

 

I'm keen as the next person for edge and fully understand it takes time to bring these products to market in a state that is worthy.

 

I would like to ask two things if I may:

 

1. How many people are holding off doing upgrades and rebuilds etc until Edge is released? I am for one.

 

2. Would ED be willing to give us 2-3 weeks notice of Edge's release so that we all can get our upgrades / rebuilds done in time?

 

 

I ask because it takes time to do these upgrades / rebuilds and on the day Edge is released I for one wish to be straight into it.

 

X9

[cichlidfan]

If you are going to do an upgrade without knowing what EDGE will require, what difference does it make if ED gives you three weeks notice?

[Rangi]

I too have been holding off on an upgrade, for over 2 years now. Another few months wont hurt to much to wait..... By then maybe the gtx 880s will be out.

[X93355]

If you are going to do an upgrade without knowing what EDGE will require, what difference does it make if ED gives you three weeks notice?

 

My thought process is that if we wait until the last moment then the latest hardware will be available and/or the prices will be less. I personally want a nice clean build without any legacy apps/settings but I want to continue playing 1.2.8 for as long as possible.

Figure 2-3 weeks notice would give enough time for those that want to do upgrades / rebuilds.

[Sephki]

I'm holding out for both EDGE and Oculus. Though, I don't quite understand the early notice. When it happens, it happens - right?

[cichlidfan]

My thought process is that if we wait until the last moment then the latest hardware will be available and/or the prices will be less. I personally want a nice clean build without any legacy apps/settings but I want to continue playing 1.2.8 for as long as possible.

Figure 2-3 weeks notice would give enough time for those that want to do upgrades / rebuilds.

 

OK, I see where you are coming from. Though I doubt ED will give the notice you would like, you can probably use the posting of a preliminary change log as an indicator. That used to precede release by a couple of weeks at least.

 

Of course, with the Open Beta program it might be harder to predict based on the change log posting.

[X93355]

I'm holding out for both EDGE and Oculus. Though, I don't quite understand the early notice. When it happens, it happens - right?

 

Yes true. I'm guessing quite a few people are holding out for upgrades. Not wanting any more notice than bare minimum helps us get those upgrades prior to release of Edge.

Are you not planning to have a clean machine for Edge?

[PFunk1606688187]

I don't really see the point in waiting. Whenever I build a computer its about getting maximum value for a given budget with the final budget being decided based on how high the performance to price curve goes with the current market and hardware iteration cycle.

 

In the end what will make DCS run best is probably going to be no different than making anything else run, with EDGE more so than before. Current DCS doesn't take full advantage of modern hardware, whereas with an engine that supports DX11 even older machines could see some boost if previously there was a bottleneck.

 

The only time I would consider holding off is when there's a major change upcoming in the market such as a new socket with a new gen of chips or a new cycle of video cards within a few weeks which could alter the price balance and so offer better value.

 

I will be getting a new PC for EDGE if only because I'm almost certain my current one will simply not be able to make the game playable in its upgraded form. I seem to be in the very very slim minority of people with a card that doesn't support DX11.

[hannibal]

i hope edge is fully optimize to use all available cpu cores..

[NineLine]

i hope edge is fully optimize to use all available cpu cores..

 

Dont know about that, but it should better utilize our GPUs and take some stress of the CPU.

[rcjonessnp175]

Sure hope your right Game_on, otherwise this is just going to be another Arma3, pretty graphics bolted ontop of a crappy non threaded engine. Lets hope thats not the case lol

[sobek]

Of course it will

 

No it won't. c0ff has clearly stated that this will not be a feature of EDGE when it will be released, but the move to EDGE will make it easier for them to, at a later point, split the whole sim up into more threads.

[shagrat]

I advise searching the forums for one or two of the "multicore" debates before ranting about crappy engines! We don't need another of these!

Simply think about the fact, that our cars still waste more than 80% of the generated power in heat and friction! Maybe it's not as easy as it seems.

[Boberro]

No it won't. c0ff has clearly stated that this will not be a feature of EDGE when it will be released, but the move to EDGE will make it easier for them to, at a later point, split the whole sim up into more threads.

 

So it is fail IMHO.

[chaos]

There are many things I'd like... Including multithreaded game-engines. Alas, it doesn't always play out that way. The approach that ED takes is a smart one, I think. Offloading cycles to the GPU is an excellent first step.

 

Baby-steps, Boberro... Baby-steps :-)

[Boberro]

Yeah, I know. But if child after years still can do only baby step it is not normal :P

[shagrat]

No it isn't. From what I understood it supports terrain preloading, that alone is a major advantage in freeing up resources. And again, what threads would you like to separate from one another? Any ideas? It is not complicated to USE multiple cores, it is difficult to find things to compute, which do not need to be synced! Fire a gun, open a new thread to calculate dispersal and trajectory, now ask the original thread when the bullets hit the target or if the target should already have been hit??? :musicwhistling:

[Flagrum]

No it isn't. From what I understood it supports terrain preloading, that alone is a major advantage in freeing up resources. And again, what threads would you like to separate from one another? Any ideas? It is not complicated to USE multiple cores, it is difficult to find things to compute, which do not need to be synced! Fire a gun, open a new thread to calculate dispersal and trajectory, now ask the original thread when the bullets hit the target or if the target should already have been hit??? :musicwhistling:

Or the other way around: it is VERY complicated to use multiple cores if you have (probably a lot of) threads that NEED to be synced (properly!). And implementing something like that is not done just by replacing the _graphics enginge_.

But EDGE is probably the best first step into that direction as it already reliefs the CPU from several tasks by handing them over to the GPU --> lesser complexity, better chance for implementing multithreading in the future.

[Kuky]

Yeah, I know. But if child after years still can do only baby step it is not normal :P

 

That depends on what kind of baby are we talking about? Human? No :D

[tintifaxl]

No it isn't. From what I understood it supports terrain preloading, that alone is a major advantage in freeing up resources. And again, what threads would you like to separate from one another? Any ideas? It is not complicated to USE multiple cores, it is difficult to find things to compute, which do not need to be synced! Fire a gun, open a new thread to calculate dispersal and trajectory, now ask the original thread when the bullets hit the target or if the target should already have been hit??? :musicwhistling:

 

I would implement an internal clock running at a 60th or 30th of a second and sync all threads to that one. So at each cycle the threads need to provide the data on an interface that are then gathered by the main thread.

 

Candidates for separate threads could be AI and LOS calculations.

 

But then what do I know about sim engines? :surrender:

[shagrat]

I would implement an internal clock running at a 60th or 30th of a second and sync all threads to that one. So at each cycle the threads need to provide the data on an interface that are then gathered by the main thread.

 

Candidates for separate threads could be AI and LOS calculations.

 

But then what do I know about sim engines? :surrender:

 

That is exactly what WON'T work!

Because all objects and their collision (LoS is a "collision detection for rays between points) need to be handled in the same thread. If not synced planes will pass through other planes or bullets etc. Now you have the same amount of calculations and simply added sync issues that need to be managed and consume additional computing power / time, resulting in even less frames per second... ;-)

 

It has been discussed over and over again. If there is an easy way to make the engine multi threaded and take advantage of multiple cores, I have the greatest trust in ED's professional developers to be capable of doing it, wouldn't you agree? :D

[Abburo]

One process over multiple threads is one of the biggest programing burdening. It is generaly avoided becasue the complexity induced is far too big.

 

Therefore, multi-core usage is based mainly for opening multiple process which have its own core. Like, my actions on CPU1 and your actions on CPU2.

 

I am pretty sure and optimistic that EDGE will have best multi CPU/GPU usage according to current resources.

 

Crhis from VEAO said recently that multi-seating will have two posible approaches and if the ED proposed one will not work then he have his own ideea how to handle it. This approach comes to underline that EDGE itself is making multi CPU/GPU opened so that multiple approaches will be possible, prone to continously improving. And this is from my point of view best approach.

Edited July 21, 2014 by Abburo