https://youtu.be/_AVp5mmc8FQ

Sup guys

Im making a new video game using 3d rad. Its going to be a physics based rock climbing game. I want to make it comical and challenging. I would like to get your thoughts on this very early prototype I have made. Its going to be simple and casual.

Cool. Laters

good luck with that, but didn't you say you wanted to try using unity a while back?

Yeah....im going to move to unity once I get  a new harddrive

https://youtu.be/x3K-TWs0Gek

Hi there!

Ive started adding my own models and trying to get a nice feel. Im going for cartoony. I'm trying to work with 3d rads realtime shadows this time.

This little video is running on an intel hd 2500 onboard gfx , running at 720 p . Captured with afterburner.

My gt 620 nvidia bombed.  :/

Hello......

I've started doing a prototype in Unity 2017 , for Rock Climber.

It took me weeks, to get the basic concept of the game working. But I'm glad it's working now.......so many tutorials........

another one bites the unity dust...

given the hardware you are using i'd be surprised if you get any better performance.

Hi thelurker.

I'm not completely sold on Unity just yet.actually trying unity at the level I'm at, made me apprecieate 3D rad even more. I realised how simple and logical 3d rad workflow is. And how badass the guys who put together 3d rad are, to have created  and managed systems of events, triggers,timers, counters, shaders, physics , and who knows what else, into an easy to use linking and relationship system. 3d rad is really a magnificent achievement on it's own.

Having said that, I am enjoying unity so far, the challenges I've faced have been so simple and laughable  I might imagine, but at least I'm on my way now.

Honestly, if I could have found a permanent fix for the inevitable crash (like a bomb on a timer) that all my games have had, while using 3d rad. I would not have even considered Unity.

will be interesting to see if unity crashes too once you import all the same assets and rewrite the game logic, if not i guess you just overloaded and out grew 3drad, but i have not had crashing issues myself.

At first I didnt have crashing. But after my compiled game would load and reload +/-   7 scenes or project files, it would crash. And there were always tell tale signs of a crash coming. The scene of project before the crash, would have weird looking shades, or other shades would be missing or incorrectly shaded, and then while trying to load the next scene of project, the game would crash.

This tell tale is what lead me to abandoning nice shaders and mostly using the default basic shader, with one other nicer shader like the water shader or immersive shader.

Another sort of fix I found was a patching program that patches any .exe to use 4 gbs of ram even when only 2 gbs of ram is allocated to .exe program.

But anyways, I enjoyed myself with 3d rad....good times spent alone dreaming of what could be....ah well.....sometimes not doing well in one subject helps you to find a different subject you might be better at. I learned to always try different paths.....unlike most of my video games progression systems......Haha Haha.

Good to hear that you're switching to Unity. The only question I have is: Why the 2017 version?

Hi Thunder. I got it from a gaming magazine from a long while back. And since my internet situation isnt great, I thought let me just mess with this version before attempting to get the latest version.

4Gb patching is useless unless your game was actually reaching 2gb ram use, which i highly doubt.
the shader issues make more sense, as your GT 620 only has 92 shader cores, and the intel only 46, while a mid range semi-modern [2017] AMD RX 570 has 2048. keep in mind nvidia cards also uses these cores for physX processing so it's no wonder too many shaders overloaded 3drad, but the same number of shaders/faces/physics objects will very likely overload unity on your pc as well, it's not the magic bullet people on here make it out to be.

you may have had some improvement by disabling physX on the GPU to force it on the CPU, but the best improvement is a better GPU, or better optimization - not even unity will do that for you. also, how many people other than you as the dev will actually load 7 levels in 1 sitting? if you can load 3 or 4 on a low-end pc without crashing that's perfectly adequate, since high-end pc's may not even crash unless you tested that as well. so don't blame 3drad - blame your hardware or game design!

The high memory usage can be reached by 3D Rad's generic memory leaks, not his game design. 3D Rad is known at least here to have memory leaks. I'm not sure about the shader cores argument though.

You can make garbage unoptimized games in both engines, but at least Unity gives you more freedom and the latest stable technology on how to do it

Ah well, whatever the reason was for the games crashing, it's not really an issue anymore. I'm just getting back to making games.

With caged photon , I also used the camera shader   glow script , which is in the sample projects, and changed the pixel shader version to 1. That helped to get caged photon  running even on intergrated gfx, like the i3 2500 I used a while back

Time for me to : q.hingejoint motor motor = motor........( I laughed so hard when I saw that in a unity tutorial ) .

Believe it or not but the 4gb patch worked, it was the only way to finish Caged photon and zero x gen in one sitting.

@nicusorn5 - prove it, exactly how much ram is leaking? is it over 2gb or even 4gb patched? then your point is pointless, as ram leakage is only an issue if available ram is depleted.

@powersupersport - yeh, but unity can't miraculously create more shader cores on low-end hardware, no matter how 'stable' it is.

@kulu - 'the 4gb patch worked' ram leaks are often caused by dodgy scripting - so was your or 3drad's code the cause? either way, an easy fix is to count the number of levels loaded per session and force a restart after a certain amount. also - how much ram do u have, how much is free before game starts and how much does 3drad actually use? likely windows is using as much of it as 3drad.

@thelurker, how do you even know whether you cause memory leaks when there's so limited documentation on how AngelScript works? We don't even know whether 3D Rad has a garbage collector. And assuming 3D Rad is aimed at beginners and "less coding" approach, how would you expect anyone to even get to the point of knowing about memory management? Even more, knowing how to get around memory management issues inside 3D Rad itself.

If one is already so deep into programming, they might as well indeed consider switching to another, better suited for programming engine (like Unity).

And yes yes, other engines won't improve your hardware of course. But they for sure could better utilize its advantages (especially new hardware features). Its explanation is as simple as "one is up-to-date, the other is outdated (8 years to be exact)".

And let's say you do get the same performance in Unity. You can still count on so many other features you wish you had in 3D Rad (ex. realtime shadows, multiple materials on one mesh, better UX in general).

Jesus Christ, why is it so hard to understand? Of course Unity is not the magic solution to fixing stupid performance or design mistakes, whether deliberate or not. But an up-to-date software is much more likely to handle such in relevant ways. Unity also has a comprehensive documentation and a bunch of online resources to guide you in the right direction.

As an added bonus, it also has much more advantages and perspective for your end product and your professional and learning development.

so limited documentation on how AngelScript works? did you not see this:
http://www.angelcode.com/angelscript/sdk/docs/manual/doc_script.html

also why would you care if you cause ram leaks or not unless you actually run out of your minimum spec of ram? do not concern yourself with such things, as there's plenty of other stuff to worry about...

in kulu's case it's shader core usage, not engine choice, and he will learn that the hard way. and even if 3drad does have shader core allocation issues, it's unlikely to affect modern mid range cards. that said, i do test on an old 32 shader core card and no issues yet but i'm not trying to shade 100's of thousands of faces with multiple shaders either, but well below 100k with maybe 5 shaders should be fine.

You'd bet that anyone would be as confused as I am when they see this link as I've never seen it being been mentioned anywhere remotely in the 3D Rad's website docs. @nicusorn5, please tell me if you know this site exists and that AngelScript uses incremental GC http://www.angelcode.com/angelscript/sdk/docs/manual/doc_overview.html. We can only hope it works right in 3D Rad.

I'll take the L on the AngelScript documentation argument. However, that doesn't invalidate any of my other points. Let's see how many ex-3D Rad people would go back to 3D Rad just because we now have the docs of AngelScript.

not all angel functions are used in 3drad, only the ones they implemented, but those that were conform to the documentation, except the dec's for arrays are slightly different.

also may want to note this before spruking GC in unity [wiki]:
'Disadvantages. Typically, garbage collection has certain disadvantages, including consuming additional resources, performance impacts, possible stalls in program execution, and incompatibility with manual resource management.'

- yep, i'm sold, switching to unity in 3...2...1...

Okay.....anyways.....yay Rockclimber!!!.......hahahahahaha......

(Long post, won't write a TLDR btw)


What the fuck? Hell no. Leaked heap memory is always a issue. I explained why myself before, but there's a StackOverflow post, since I'm not bothering anymore:
(As a argument 3D Rad's memory leaks are repetitive). See [1] and [2]



Yes, I know that this documentation website exists. I even used it myself when I found out that you can create your own classes in 3D Rad's scripting engine. But I didn't know about AngelScripts
s GC. But as far as I've looked, AngelScript does have a GC . Anyway, we know that 3D Rad's memory leaks cannot be fixed by `fixing` our scripts. I don't even see how you can leak memory in AngelScript, since it is a scripting language, and as far as I'm concerned, all memory allocations (heap or stack) are both fully automatic. ( global variables on the heap and local variables on the stack)


Yes, occasional GC collections are known to freeze games and stuff. Remember that Unity is made for all kinds of developers (with and no experience in programming). So of course they would choose C#'s GC instead of forcing the users to allocate the memory themselves.
    As a measure for that, C# game devs write "GC-friendly" code (as many as less class instance creations), using ECS (entity component system), or AOS(array of structures) instead of OOP to limit cache misses by making the memory as linear as possible. By the way, experimental Unity builds are using the ECS paradigm for their object storage.

Sources (since some of you want proofs):
[1]
https://classdev.net/page/thread/1IlUJ3_D6i787H6C/3drad-bugs-revealed (Second page, kulu's first post).
[2]
https://stackoverflow.com/questions/1719267/do-small-memory-leaks-matter-anymore
[Other proof]
https://learn.unity.com/tutorial/entity-component-system
https://en.wikipedia.org/wiki/CPU_cache#Cache_miss
https://www.quora.com/Why-is-object-oriented-programming-OOP-slower-than-procedural
https://www.linkedin.com/pulse/ecs-beginning-end-oop-sander-mertens

quote athanatos: 'I don't even see how you can leak memory in AngelScript....all memory allocations are fully automatic'
wrong, you didn't read the documentation on iShaderTextureDestroy did you? same goes for iObjectImpostersDestroy, and iDLLUnload, ram for these is only released manually by the user, or on program exit.

and you still haven't proved how much time and ram use caused the crash, nor exactly how many bytes were leaking or from where, from what i can see over 1hr monitoring private/virtual bytes in perfmon, it might be 10mb per hour, about the same as my AMD driver service maybe leaks, not to mention firefox, which is well known to leak 1000's of mb [or is it a false positive because it is actually just caching data directly proportional to number of tabs open] yeah, like i must switch to unity to free 10mb/hr on a game using 512 or even 1024mb on a  8192mb system are you serious!?

kulu said physx dll had errors, so there's no proof 3drad itself is leaking at all, but some of nvidia's dll's can be upgraded, others bypassed by forcing cpu physX. also errors in Intel Inspector are not classified as critical, crash causing bugs, because if still plenty of free ram during typical use they are usually harmless and will not cause a crash. and even your stack link has comments that support me saying leaks can be fixed by simply timed/counted restarting of the software, which is all kulu needed to do really.

Not wrong, this is not AngelScript's fault. These are the functions that 3D Rad exposes to the AngelScript interface.



Not to mention kulu's crashes after loading/unloading ~7 projects? Guess I'm going to have to do some tests myself.

i already set up a test multi level project where the problem is worse as kulu said, private ram use on 1st load is as follows:
main menu=80mb
game level=270mb [virtual ram=360mb]

switching between both adds 50mb each time so after 5 level+5 menu loads i got:
main menu=330mb
game level=520mb [virtual ram=920mb]
yeh, that's worse than i thought, but still will take a while to crash my pc, even if my game triples to 1gb, but need more tests to see if level load increases from 50mb but so far it added 50mb even to/from the 80mb menu with only 1 sprite/skybox.