Friday Update, June 01, 2012

[zapatista]

Originally Posted by Thee_oddball

those are caused by calls to non native library's (C++) which cause interop which causes ...stutters, the only way i know to fix that is to rewrite all the C++ libs to in .NET.

meaning what ? a few edits in some lines of code that are critical for this function, or months of rewriting large chunks of code ?

[Blackdog_kt]

Originally Posted by zapatista

meaning what ? a few edits in some lines of code that are critical for this function, or months of rewriting large chunks of code ?

Depends on the size of the libraries. It's definitely not a copy/paste thing though and it goes into more than just "translating" the syntax from one language to another.

For example, say i have a small game written in Java or C# on my PC. These languages have what is called an automatic garbage collector: any entity that has served its purpose and not being used anymore is automatically purged from the PC's memory.

Now let's say i want to port my little game to objective C so that i can sell it on the apple store and make money off people with iPhones. I have to comb through the code and manually tell it exactly what and when to purge from memory, because it's not automatic.

The behaviour of certain commands can also complicate things, in one language a command does something and exits that block of code, in another language it might do things in a different sequence, etc.

All in all there is some complexity involved, so the deciding factor here is the amount of code to be "translated" between languages.

[Thr0tt]

lol, here's some new aircraft from a new game we are working on, oh btw yeah were working on this patch thing.

Nice, all I can say is the day 1 purchasers for your product is going to be pretty non-existent, its a shame as it could be the best game out there but will be battling from the start a negative vibe.

[Continu0]

Originally Posted by Blackdog_kt

Depends on the size of the libraries. It's definitely not a copy/paste thing though and it goes into more than just "translating" the syntax from one language to another.

For example, say i have a small game written in Java or C# on my PC. These languages have what is called an automatic garbage collector: any entity that has served its purpose and not being used anymore is automatically purged from the PC's memory.

Now let's say i want to port my little game to objective C so that i can sell it on the apple store and make money off people with iPhones. I have to comb through the code and manually tell it exactly what and when to purge from memory, because it's not automatic.

The behaviour of certain commands can also complicate things, in one language a command does something and exits that block of code, in another language it might do things in a different sequence, etc.

All in all there is some complexity involved, so the deciding factor here is the amount of code to be "translated" between languages.

Thank you!

[Thee_oddball]

Originally Posted by zapatista

meaning what ? a few edits in some lines of code that are critical for this function, or months of rewriting large chunks of code ?

they would have to write the tree's in C# to get rid of the interop (performance penalty)

FYI the game uses something called "speed tree's" http://www.speedtree.com/

[Osprey]

Originally Posted by Thee_oddball

those are caused by calls to non native library's (C++) which cause interop which causes ...stutters, the only way i know to fix that is to rewrite all the C++ libs to in .NET.

C++ is native, not non native. What makes you think there are perf problem between the layers anyway? If so the answer would be to optimise that, not convert.

[Thee_oddball]

Originally Posted by Osprey

C++ is native, not non native. What makes you think there are perf problem between the layers anyway? If so the answer would be to optimise that, not convert.

the core game is written in .NET and the tress are a C++ DLL hence the interop, and What makes you think there are perf problem between the layers anyway? is the fact that you hear that people have high FPS but still have stutters which gets fixed by turning OFF the tree's should be your first indicator that interop is causing a performance hit.

optimization? the game was in dev for 5 years before release...it came out of the gate with no doc, no dedicated server soft and very little content...at the very least it should have been optimised at release.

[adonys]

have it ever crossed your mind that it might be because of a bad implementation of the speed tree library?!!

[Buchon]

Originally Posted by Thee_oddball

the core game is written in .NET and the tress are a C++ DLL hence the interop, and What makes you think there are perf problem between the layers anyway? is the fact that you hear that people have high FPS but still have stutters which gets fixed by turning OFF the tree's should be your first indicator that interop is causing a performance hit.

optimization? the game was in dev for 5 years before release...it came out of the gate with no doc, no dedicated server soft and very little content...at the very least it should have been optimised at release.

Ohh ... this is an interesting theory, I did not understand you at first but now is more clear.

But ... I recall that this hick-ups (no fps affected) are not in the old build, they was micro-freezes (fixed in the new build).

But you are right ... it can be a interop problem but I think that can be caused by the issue of left the tree and grass code aside of the new build that is in the Alpha patch, so the ddl dependence is high.

Has is written in the know-issues list in BlackSix patch post, the tree and grass optimizations was not complete but it will be in a few days.

Anyway we´ll see this soon



By the way, to clear misunderstandings in the DX9/10/11 debate :

Of course you can made a game with DX9/10/11 support but what I was saying is that by doing it you´ll not fully use the DX11/10 possibilities.

That´s it.

Cheers

[mazex]

Originally Posted by Thr0tt

lol, here's some new aircraft from a new game we are working on, oh btw yeah were working on this patch thing.

Nice, all I can say is the day 1 purchasers for your product is going to be pretty non-existent, its a shame as it could be the best game out there but will be battling from the start a negative vibe.

Well, as the new game has a goal to reach alpha in 2013 and the ongoing development on the base engine will benefit CloD it would surprise me if the game is not rather OK by then.

The latest alpha/beta is actually rather OK, but we are so "tainted" that we don't see it. I think it is what version 1.0 of CloD should have been... Working OK on high end rigs and lots of fixable bugs

/mazex