Graphics engine from original IL2 utilised in CloD. So what hope for BoM?

[LoBiSoMeM]

Originally Posted by ZaltysZ

Object density and performance have linear relation. Drawing distance and performance have cubic relation. Doubling drawing distance requires not 2 times more, but 4 times more computing power. This means that drawing distance will always be a pain, unless you want to sacrifice much density and complexity of objects. Other games simply add fog outside of the bubble, so objects don't pop up, but appear out of fog instead. Cosmetically it looks nicer, but you simply loose ability to have different drawing distances for different objects. I would personally choose to live with popups, than with fog, which masks rivers, hills at distance and make navigation uncomfortable.

Perfect!

We just need that important buildings (like ground objectives, bases buildings, radar stations, etc) have huge drawning distance. In FMB for the sequel maybe we can set this and bomber pilots can have a easier life... Just a suggestion!

There is simply no time, or even more importantly money, to build a new game engine.

[LoBiSoMeM]

Originally Posted by Ma233e

There is simply no time, or even more importantly money, to build a new game engine.

And the game engine now is good. Regards "drawning buble" we maybe need some tweaks in view distance of some important objects, just that. Wll be good for navigation and level bombing runs.

[Stublerone]

Originally Posted by Ma233e

There is simply no time, or even more importantly money, to build a new game engine.

There is no need to rewrite engine as they already did. And it is a problem of the of the technique together with the high viewing rang. Noone on this earth could make it sufficient with this method and the given parameters.

We are talking about an open world, large maps, big texture and clear and ultra long viewing range. Noone ever get this work as you all intend. Simply not possible with our "slow" machines to compute all the workload needed to get a really good result you are looking for. Impossible with all the slow interfaces between the components. Not to speak of a completely revolution in using multicore cpu and multicore graphics (sli has to function totally different with a much more effective interface between the cards). Also fasterssd, fast as ddr3 at least and change ddr3 ram to ddr5 vram speed and get the right and fast buses between them. If the pc is working like this, you could PERHAPS be able to generate a sufficiently working stream engine game with some free capacities for graphics evolutions.

So: really difficult technically!

[zapatista]

Originally Posted by Stublerone

There is no need to rewrite engine as they already did. And it is a problem of the of the technique together with the high viewing rang. Noone on this earth could make it sufficient with this method and the given parameters.

We are talking about an open world, large maps, big texture and clear and ultra long viewing range. Noone ever get this work as you all intend. Simply not possible with our "slow" machines to compute all the workload needed to get a really good result you are looking for. Impossible with all the slow interfaces between the components. Not to speak of a completely revolution in using multicore cpu and multicore graphics (sli has to function totally different with a much more effective interface between the cards). Also fasterssd, fast as ddr3 at least and change ddr3 ram to ddr5 vram speed and get the right and fast buses between them. If the pc is working like this, you could PERHAPS be able to generate a sufficiently working stream engine game with some free capacities for graphics evolutions.

So: really difficult technically!

one further possible avenue of increased performance for game programers is to directly access/write-to GPU functions (instead of using intermediary languages as they do now, which need 1 or several steps between the hardware function and the game code written), but i am not aware of any current game that does so yet. from what i read, this will allow a significant further step forward in performance nd some upcoming FPS are sarting to use these methods afaik. future more efficient game code writing for flightsims is also possible even using current harrdware, but i agree that right now given the current method of writing these games, performance for flightsims is very much strangled by what the current hardware can perform, and it is a game genre that is pushing the boundaries of what is possible.

[ZaltysZ]

Originally Posted by LoBiSoMeM

Perfect!

We just need that important buildings (like ground objectives, bases buildings, radar stations, etc) have huge drawning distance. In FMB for the sequel maybe we can set this and bomber pilots can have a easier life... Just a suggestion!

Bomber guys would be happy even with such simple solution as separate view distance slider for bombsight mode. It is better to do slight adjustements at 15FPS and see your target in advance, than to have it poping up when it is already too late for corrections.

[tintifaxl]

Originally Posted by Stublerone

There is no need to rewrite engine as they already did. And it is a problem of the of the technique together with the high viewing rang. Noone on this earth could make it sufficient with this method and the given parameters.
...

You contradict yourself in the next sentence. So a rewrite of the engine with support for multicore cpu's and sli gpu's is necessary to achieve better results.

Why do I have such a fast and expensive system? So developers can still not support multicore cpu's, multiple gpu's and use slow execution programming languages like C#, that have a huge overhead when using the Direct3D API?

[LoBiSoMeM]

Originally Posted by ZaltysZ

Bomber guys would be happy even with such simple solution as separate view distance slider for bombsight mode. It is better to do slight adjustements at 15FPS and see your target in advance, than to have it poping up when it is already too late for corrections.

Yes, maybe the two things can be put together in BoM: bigger draw distance for important objects defined by mission builders AND general bigger view distance for bombsight.

[Stublerone]

Originally Posted by tintifaxl

You contradict yourself in the next sentence. So a rewrite of the engine with support for multicore cpu's and sli gpu's is necessary to achieve better results.

Why do I have such a fast and expensive system? So developers can still not support multicore cpu's, multiple gpu's and use slow execution programming languages like C#, that have a huge overhead when using the Direct3D API?

You are right with second. But what you claimed in your first part: I just wanted to say that sli is bad, multicore cpu is bad as long its buses, interfaces and the way they are handled gets improved. Sli for example do not us both cards capabilities 100% and it is even worse with its vram just 100%, where it should be 200% with 2 cards. It is just using the amount of 1 x vram, which is rediculous especially when upgrading a 1gb vram card. You can directly take the money as toilet paper, because it does not help you get rid of the problem of texture load in clod. You can buy 50 cards and generate 3000 fps in clod without getting rid of the last hick ups.

So sli needs rework and another implementation. 3 cards for 3 monitors will do 3 parts of thepicture. 1 for every monitor and without loss of the potential vram. -> That would be an sli, which I directly buy, no matter how much bucks. But you will perhaps never see such an upgrade politics from the manufacturer, because he cannot sell his top product for gaming, if the people can upgrade sufficiently. Why should he do that? He earns not as much and perhaps needs to produce lower end or nearly obsolete cards a longer time. Not a sufficient model for him. So he will stick with the politics, which is not doing him any harm, but us! Simple business. Add the fact to use prematerials that brake after 4 years and you are a rich company! And noone really can do anything as long as it is needed and as long as there is a market for it.

As long as the coms between the hardware parts are just evolving as now, we will always have such a problem. Usb2, usb3, sata 300, sata 600, thunderbolt, firewire, .... All crap to the possibilities, that they could integrate. The hardware is to far away from each other