A new simulator engine will meet customers' most crucial demands!!!

[Marcio]

The physics involved in the parameters of flight, ballistics and damage, the variety of aircraft and vehicles, the historical care in the campaigns and the details of cities and landscapes are highlights of Il-2 Great Battles line of air and armored vehicles simulation.

 

However, the introduction of the Me 109 G 06 AS and the upcoming arrival of the high altitude interceptor, Ta 152 exposes a considerable gap for a WWII simulator; the absence of classic four-engine USAAF and RAF aircraft even in the form of A.I.. Even the current status of the simulator exposes some bottlenecks, notably in missions like Bomb Beachhead and Patrol Beachhead on the beaches of Normandy with Mullberry ports, where the simulator drops in performance and becomes extremely slow, even on a gaming PC with an Intel processor Core i7 with 32 Gb of RAM from Corsair, Sound Blaster sound card and NVidea RTX 2600 Super GPU. I believe that the most appropriate solution leans much more towards optimizing the simulator's engine than upgrading the components of the gamer PC and I say this based on the superior performance of my gamer PC in another simulator of similar carat. I believe that a new engine should focus, above all, on the use of 2 or, preferably, more independent processor cores, each one corresponding to a thread of execution. An upgrade to a development interface capable of making communication between software and hardware even easier like migrating from DirectX XI to XII would help even more.

 

The improvement in simulator performance much more due to the optimization of its engine than due to an upgrade of components of a gamer PC, in addition to relieving the simulator user, freeing him from purchasing very expensive hardware and making the simulator inclusive to a larger number of people, it would meet a series of customer demands. From more desired demands such as career mode and online battles including a greater number of aircraft, tanks and ships, the presence of four-engine aircraft and large air-naval battles including aircraft carriers, battleships and cruisers; ranging from demands such as the possibility of landing, repairing, refueling and rearming aircraft at any airfield or allied aircraft carrier to demands that are not so crucial but which, without a doubt, would be an added attraction for the simulator, such as larger maps, maritime, fluvial, railway, road and air transport traffic of civil vehicles, presence of birds and livestock, greater number of plant species, enhancement of autumnal foliage in temperate forests, etc.

[AndyJWest]

ChatGPT?

[Trooper117]

Here we go...  

[R33GZ]

2 hours ago, AndyJWest said:

ChatGPT?

You jest... but if you got ChatGPT to generate a statement based on user requests, I think that is pretty much what it would spit out ?

 

I had to look up fluvial - it means river traffic ?

[AEthelraedUnraed]

6 hours ago, Marcio said:

the presence of four-engine aircraft

..is not related to the engine; the Devs have said multiple times that it's perfectly possible to create 4-engined aircraft but that the relative amount of work they'd have to put into (flight) modeling makes them unattractive commercially.

 

6 hours ago, Marcio said:

aircraft carriers, battleships and cruisers

Similar here, the Devs have long said that the problem with aircaft carriers is the amount of work that goes into them, as well as into believable physics for them.

 

6 hours ago, Marcio said:

career mode

Not sure what you're getting at? IL2 has had a career mode for ages.

 

6 hours ago, Marcio said:

the possibility of landing, repairing, refueling and rearming aircraft at any airfield

Likewise, this has been possible for years, as long as the mission writer has enabled it.

 

6 hours ago, Marcio said:

larger maps

Again, the required development effort is the problem, not the engine.*

 

* fun fact: map locations are stored in a 64bit float format which, if my quick calculation is right, is enough to store every location on the earth with about 10 nanometer accuracy.

 

7 hours ago, Marcio said:

maritime, fluvial, railway, road and air transport traffic of civil vehicles

Already possible, at the discretion of a mission writer. There are already civilian ships and river barges available, the C-47, Ju-52 and U-2VS were widely used for civilian aviation (which was pretty much non-existant on all of the current maps during WW2), and enough of the land vehicles are passable as civilian ones.

 

7 hours ago, Marcio said:

greater number of plant species

Although there is a limit of 8 tree models per forest type, there is no such limit on the whole map beyond a maximum of 15 forest types.

 

7 hours ago, Marcio said:

enhancement of autumnal foliage in temperate forests

No idea what you mean by this. The autumn forests already look different than the summer/winter ones.

[Yogiflight]

8 hours ago, AEthelraedUnraed said:

The autumn forests already look different than the summer/winter ones.

The issue isn't the autumn forests, but the non existing winter forests on the Moscow map (and others?). With snow all around and very low temperatures you should have white trees, not to forget deciduous should be without leaves in winter.

 

8 hours ago, AEthelraedUnraed said:

the Devs have said multiple times that it's perfectly possible to create 4-engined aircraft

yes, but they couldn't fly in reasonable numbers for intercept missions with the Bf 109 G6AS or the Me262

 

8 hours ago, AEthelraedUnraed said:

15 hours ago, Marcio said:

larger maps

Again, the required development effort is the problem, not the engine.*

well, when I see how the towns get loaded through the flight on the Rhineland map, I would guess the engine can't take any more

 

8 hours ago, AEthelraedUnraed said:

15 hours ago, Marcio said:

aircraft carriers, battleships and cruisers

Similar here, the Devs have long said that the problem with aircaft carriers is the amount of work that goes into them, as well as into believable physics for them.

in theory sure, but not with the amount of AA guns, aircraft carriers and battleships were carrying, at least not as long as AA guns shoot with one AI for each AA gun and not one AI for one or more batteries of AA guns of the same caliber

[senseispcc]

I fly only "not" online and all is "OK" . 

[Marcio]

16 hours ago, AndyJWest said:

ChatGPT?

ChatGPT?!

I expected a more technical and enlightening answer!

This is just a constructive suggestion to improve the simulator's engine in order to correct the slowness, notably in missions like Bomb Beachhead and Patrol Beachhead on the beaches of Normandy with Mullberry ports. This should not happen on a PC with an Intel Core i7 processor with 32 Gb of RAM from Corsair, Sound Blaster sound card and NVidea RTX 2600 Super GPU. I say this based on the superior performance of my PC in another simulator of a similar level, where this slowdown does not happen, even in complex missions. I suggest that a new engine should focus, above all, on the use of 2 or, preferably, more independent processor cores, each one corresponding to a thread of execution. An upgrade to a development interface capable of making communication between software and hardware even easier like migrating from DirectX XI to XII would help even more.

The improvement in simulator performance much more due to the optimization of its engine than due to an upgrade of PC game components, in addition to relieving the simulator user, freeing him from purchasing very expensive hardware and making the simulator inclusive to a larger number of people, it would meet a series of customer demands. Unsurprisingly, many customers want, for example, the presence of four-engine bombers and large air-naval battles including aircraft carriers, battleships and cruisers.

[AndyJWest]

As has already been noted, this has all been discussed before. 

 

If you want a 'technical' answer, you are unlikely to receive anything much from those in a position to know, until the developers reveal what their future plans are. And in any case, the constraints are more likely to be financial than technical. If the developers had unlimited resources I'm quite sure they could deliver a lot more than they currently do. Development of a new game engine costs money - a lot of it.

[Marcio]

14 hours ago, R33GZ said:

You jest... but if you got ChatGPT to generate a statement based on user requests, I think that is pretty much what it would spit out ?

 

I had to look up fluvial - it means river traffic ?

 

Am I joking?

 

I'm going to pretend I didn't read that and thought that's who you are!

 

I just suggested a new engine focused mainly on the use of 2 or, preferably, more independent processor cores, each one corresponding to a thread of execution and a migration from DirectX XI to XII in order to improve the simulator's performance and meet the desired by a considerable number of customers.

 

Any other suggestions for improving the simulator?

[Marcio]

28 minutes ago, AndyJWest said:

As has already been noted, this has all been discussed before. 

 

If you want a 'technical' answer, you are unlikely to receive anything much from those in a position to know, until the developers reveal what their future plans are. And in any case, the constraints are more likely to be financial than technical. If the developers had unlimited resources I'm quite sure they could deliver a lot more than they currently do. Development of a new game engine costs money - a lot of it.

The simulator is good, but it can be improved without a colossal amount of financial resources.

With regard to the engine, there is no need to reinvent the wheel! Top-notch simulators with four-engine bombers and aircraft carriers already exist!

Possibly fairly trivial solutions like including more processor cores in the engine's workings can help.
Anyway, I like the simulator and I believe that if demands such as naval air battles including aircraft carriers and cruisers and the presence of four-engine bombers, even if operating as A.I. are met, the number of customers will increase a lot.

[Panzerlang]

As much as I'd like 4-motor bombers to shoot at I'd prefer a more complex damage-model (undercarriage damage would be a very nice addition).

[AndyJWest]

1 hour ago, Marcio said:

...fairly trivial solutions like including more processor cores in the engine's workings can help...

 

Creating multithreading software with the level of complexity an air combat simulator involves is not 'trivial'.

[R33GZ]

3 hours ago, Marcio said:

 

Am I joking?

 

I'm going to pretend I didn't read that and thought that's who you are!

 

I just suggested a new engine focused mainly on the use of 2 or, preferably, more independent processor cores, each one corresponding to a thread of execution and a migration from DirectX XI to XII in order to improve the simulator's performance and meet the desired by a considerable number of customers.

 

Any other suggestions for improving the simulator?

Keep your pants on mate, I was replying to Andyjwest. No offense intended if that's where you ended up. Exclamation mark.

[Marcio]

58 minutes ago, R33GZ said:

Keep your pants on mate, I was replying to Andyjwest. No offense intended if that's where you ended up. Exclamation mark.

     True!

     You were replying to Andyjwest and what you answered is, no doubt, very appropriate!
"... if you got ChatGPT to generate a statement based on user requests, I think that is pretty much what it would spit out". 

     Unsurprisingly, many users want, for example, the presence of four-engine bombers and large air-naval battles including aircraft carriers, battleships and cruisers.

 

     A user suggest improvements to the simulator engine and a founder responds with a "ChatGPT?"    

 

     From a founder the user expects a technical answer and not a disdain!

 

     Sorry!!!

     Earlier, when I was returning from a scubadive, the small vessel I was on, sailed through a windy region and the spray of brine left my vision blurry! Besides English not being my native language, when I read your reply on a small smartphone screen, my eyes were blurry!

[R33GZ]

4 hours ago, Marcio said:

     True!

     You were replying to Andyjwest and what you answered is, no doubt, very appropriate!
"... if you got ChatGPT to generate a statement based on user requests, I think that is pretty much what it would spit out". 

     Unsurprisingly, many users want, for example, the presence of four-engine bombers and large air-naval battles including aircraft carriers, battleships and cruisers.

 

     A user suggest improvements to the simulator engine and a founder responds with a "ChatGPT?"    

 

     From a founder the user expects a technical answer and not a disdain!

 

     Sorry!!!

     Earlier, when I was returning from a scubadive, the small vessel I was on, sailed through a windy region and the spray of brine left my vision blurry! Besides English not being my native language, when I read your reply on a small smartphone screen, my eyes were blurry!

All good man. For what it's worth, your English is very good... also, I think 'founder' just means someone who invested back in the early days of GB and is recognized as such. They're not 'company men' They can get a bit, er.... salty,  when topics that are well trodden get brought back to life though ?

[JG27_Steini]

12 hours ago, Marcio said:

The simulator is good, but it can be improved without a colossal amount of financial resources.

With regard to the engine, there is no need to reinvent the wheel! Top-notch simulators with four-engine bombers and aircraft carriers already exist!

Possibly fairly trivial solutions like including more processor cores in the engine's workings can help.
Anyway, I like the simulator and I believe that if demands such as naval air battles including aircraft carriers and cruisers and the presence of four-engine bombers, even if operating as A.I. are met, the number of customers will increase a lot.

 

The current engine run only at 1 CPU core. Having more cores than usual does not bring any positive result, it even may slow down the game if the core single speed is lower than other CPU have. I am sure the new engine will have a multicore support. Otherwise it would not fit the demand of a future ready engine. In future the single core performance wont improve much, but more multicore CPU will appear. New games have to take all performance they can get and using all cores is a must have for a modern game. In the past, they missed that chance, they will have to in the future or it wont last much longer.

[Trooper117]

[76IAP-Black]

[AEthelraedUnraed]

19 hours ago, Yogiflight said:

The issue isn't the autumn forests, but the non existing winter forests on the Moscow map (and others?). With snow all around and very low temperatures you should have white trees, not to forget deciduous should be without leaves in winter.

I don't think that's what the OP meant; autumn != winter. Besides, this is a texturing issue, not an engine one

 

20 hours ago, Yogiflight said:

well, when I see how the towns get loaded through the flight on the Rhineland map, I would guess the engine can't take any more

There is a limit to how many buildings can be shown while still having reasonable performance, yes, and an engine upgrade might be able to improve that, as well as the loading time for those towns. That's not related to the size of the map though, as cities are loaded piecemeal (that's actually what you're seeing).

 

18 hours ago, Marcio said:

I just suggested a new engine focused mainly on the use of 2 or, preferably, more independent processor cores, each one corresponding to a thread of execution

5 hours ago, JG27_Steini said:

The current engine run only at 1 CPU core.

IL2 already does run on multiple cores You can easily check this:

Just a 4 vs 4 quick mission on the Lapino map. As you can see, there are 85 different threads, of which in this case 9 use any kind of significant CPU time (I have to admit I got lucky pressing the print screen button; usually it hovers around six) and a similar amount of threads use a significant amount of CPU cycles.

 

Unless the Devs have manually specified a certain core affinity (which is unlikely), Windows will automatically distribute the threads around all available cores.

(Also, there isn't really any difference between developing for 2 or for 20 cores. You either use significant threading, in which case your program runs well with more cores, or you don't, in which case more cores won't generally help.)

 

18 hours ago, Marcio said:

a migration from DirectX XI to XII in order to improve the simulator's performance

While DX12 is said to improve the CPU overhead, it's still a performance improvement in the graphics part and that's not where IL2's main bottleneck is. It'll help, definitely, but I doubt that migrating from DX11 to 12 makes any huge performance impact on how many aircraft you can use simultaneously.

Also, any such migration is anything but trivial, from what I've read about DX12.

 

19 hours ago, Marcio said:

This is just a constructive suggestion to improve the simulator's engine

In all fairness, I don't think suggestions like these are really useful or constructive. Half of the things you mention already exist in the game in some form or another, and about the other half, everyone who has participated in this thread (including myself) knows too little combined with this specific usecase to say anything useful. The fact is that only the Devs know how their code looks like, and where the main bottlenecks are. Things like upgrading to DX12 or using more threads sound nice, but if done improperly may bring more processing overhead than they intend to solve, or implementing them may cost more time/money than the Devs can spend.

 

I'm 100% sure that the Devs have looked at different ways to improve IL2's performance, including the suggestions you've made. If it's feasible and brings a large enough performance improvement, it will be done in the new "project"; if it isn't done in the new "project", it apparently wasn't feasible or didn't actually improve things much.

[FuriousMeow]

11 hours ago, JG27_Steini said:

 

The current engine run only at 1 CPU core. Having more cores than usual does not bring any positive result, it even may slow down the game if the core single speed is lower than other CPU have. I am sure the new engine will have a multicore support. Otherwise it would not fit the demand of a future ready engine. In future the single core performance wont improve much, but more multicore CPU will appear. New games have to take all performance they can get and using all cores is a must have for a modern game. In the past, they missed that chance, they will have to in the future or it wont last much longer.

 

Complete bunk.

[SharpeXB]

On 8/22/2023 at 11:59 PM, AndyJWest said:

ChatGPT?

I was thinking the same thing! ?