I love text code, and logic code,
logic just represents code, so in theory a logic brick is a script with variables, but they are more “hardwired”
The baking is still not happening properly for some reason. The automatic UV unwrap could be part of the problem. Plus , I think that the bake margin is not working as it should.
Attachments
Tile Map Bake.blend (1.6 MB)
Problem is , I’ve seen some extremely complicated flow-graphs in Cry Engine that were impossible to figure out. How is that making it easier … I think the idea for node editors / flow graphs was to make scripting easier … That’s not always happening.
As I have said before you can only make it so much easier before it isn’t programming and therefore not functional. The reason that we have jobs in the world is because people still have to train to perform different tasks. That’s a fact of life and no different here. The reason that there is a game design industry is because it’s hard.
Sent from my Nexus 5 using Tapatalk
You are right! But what is a “full game engine”?
By full game engine, I’m talking about a dedicated platform. Blender is a modeling and all around creation platform, but it’s not dedicated to any one activity, and BGE is just an extension of that. A full game engine would be like CryEngine, Frostbite, Unreal, Unity, SheVa, etc.
I see, thx.
LOL, if you want https://www.youtube.com/watch?v=_F-hwyu9bcU&feature=youtu.be
then use CRYENGINE USD$9.90 amonth ONLY.
AFAIK, Blender Game Engine will NEVER be able to achieve it not even AAA.
Polycounts are way too limited in Blender Game Engine.
Go use CRYENGINE for this type of things.
www.CryDev.net
Blender is ONLY good for modeling, export your model using CryBlend(they have FBX converter called RC).
have you looked at poly limits since 2.7x?
I have to put over 200,000 on my msi afterburner nvidia 210 before I lose a 3fps,
I think the only things needed appear to be in the works anyway,
anything you can
hardware skinned armatures
Differed lighting
etc.
Look at what that Kupoman and the amazing moogle are doing.
Will BGE have this capabilities some day?
Short answer: Never
Long answer : Never until the end of the universe. Sad but true :spin:
Go to www.unrealengine.com pay $19,00 and be happy
Short answer: Never
Long answer : Never until the end of the universe. Sad but true :spin:
I don’t get your sentiment,
there is native code, that is on the cpu, that is slow,
Move what you can to gpu,
What “magic” does ue4 have we don’t?
write it?
there are some python issues, just use lil tiny pythons and keep gil happy 
What problem?
UE4 is not written in Python, it’s written in a C variant of some sort, which makes it more efficient. Python is great, but all of BGE’s functions are written in it, which is much more taxing on your machine. They also probably created their own language to auto correct any memory leaks, and anything in Python does have a tendency to leak over extended periods of time. BGE is infamous in this perspective-play in it for more than an hour and your RAM starts to bleed.
If you were to go back and write more in Blender’s C heritage, then you could get some much more RAM/CPU friendly results.
I’d love to see more done in BGE being supported by the GPU, instead of the CPU, but that really doesn’t fix the issue. While much more powerful, the GPU will overload too. Everything has a limit. :spin:
logic bricks are C++,
Yes, but you write in Python, do you not? Every script you write for BGE is just a specified logic brick. I’m fairly certain that Blender also compiles Python to C++, then C++ to C. No matter what, it’s being going through other languages, effectively slowing down any execution.
What problem?
The so much longer answer:
I worked with BGE for the last 5 years with payed works and teaching it and break my heart to say it but each day BGE is closer to death. BF give it up ( if not officially, did it by actions ). The minor staff of developers are working in some other things (Dfelinto in bake, Breach in cyles, Campbell in all other things) today we have just Kupoman and moguri working in it without any payment for it so what is the future of BGE? Ton promised to look at it in 2.8x and turn it part of blender as a playground for simulations and such kind of thing so it’s the official end of the BGE as we know and i think it’s a good thing cause if BGE don’t deserve any attention as a Game Engine maybe it can bee a good some else.
In the Technical Point BGE is worse even to Unity3D free, JMonkey and others
-
BGE is based in OpenGL 2.1 and it’s old ( 8 years ) it’s designed to PS2 with little improvements until now
-
Inefficient internal Scene Management ( Try push 200,000 objects with 1 triangle to see your cpu/gpu burning out )
-
Lack of HDR and Deferred rendering pipeline
-
Python is not the bottleneck, single thread processing is.
-
Inefficient Culling system
-
We don’t have even a Optimized Material System ( the code that we made in Material Editor and Node is messy as the hell, download my add-on to export GLSL code and see it )
For all this points we can’t just fix it in some way. It’s the same that do a completely new game engine, actually some guys are just doing it in Gamekit project. ( what i personally don’t expect too much )
My position about any tool is use it as a tool and not for love. I don’t want to love my pencil and if it’s broken and a new one is on the market and it’s so much better and accessible as my current so i think is a good thing to put it in the garbage and get the new and better one. I did it with BGE. I don’t think it’s easy cause we actually are emotional people and we spend part of our lives working on it, learning and hopping that it someday will receive some attention but one day you just see that we have nothing more to expect about it. So you can dream with the day that BGE will grow up and become a great game engine or you can just pay $19,00 for once, cancel the subscription and have a AAA game engine in your desktop today or even download a free version of the Unity3D ( best indie Game Engine ). Just think about it.
So - openGl, how hard to go to current?
Objects - can we push this into cuda/gpu?
I don’t use a lot of python,(for speed) if logic bricks were setup for sensors to have the option to pass a variable to a property, you
don’t need near as much python.
optimized material settings-how long of a recode for this? what about nodes?
This is a great explanation of BGE’s current situation.
BGE is pretty much shut down, and rarely sees any work done to it, but Ton Roosendaal did promise us a 3.0 Interactive mode with full access to Blender facilities.
The main problem is not that it’s (necessarily) outdated, but two reasons:
- It is an extension of Blender, not a game engine. It’s what is called an Interactive renderer. 3.0 BGE will be renamed to be “Interactive mode” which is much more fitting.
- It’s written in Python. Python is great, don’t get me wrong, and I love it, but it’s great for rapid development, not performance. It’s extremely advanced, even moreso than C# in my opinion, and just doesn’t have the ability to run a powerful engine. You need a low level language for that, which takes much longer, but will have the best performance by far.
BGE also has extremely poor memory management. This is mostly due to the fact that it’s composed of Python by mostly drifters. As far as I’m aware, no one has ever been paid for their work on BGE. It’s really just a collection of code that was never revised. Most of the guys who wrote it only wrote Python, and honestly did good work, but if you only know Python, memory management isn’t going to be a strong suit, because you don’t know how the language is actually working.
I’m not trying to insult BGE at all, because I truly love it, but it’s just not a real engine.
As for Unity, I could argue that being the best indie game engine…
What about with a whole team of coders?
That honestly only works so far. I rarely use Python, (most of my philosophy on language is adopted) but that doesn’t stop me from having to manage my games very strictly.
Martins Upitis makes some of the most wonderful effects I’ve ever seen in the BGE, and he rarely uses logic bricks except to plug in his scripts. All his work also flows perfectly on basic computers.
There is validity in your idea, but Python is good for plugin scripts. It’s when a program is composed of Python that it becomes a problem. If you had the time to work on it, you could always recreate an external BGE using a directly C derived language, but I wouldn’t do it personally.