What does this class hierarchy have to do that it needs this kind of logic and complexity?
[QUOTE=marvincmarvin;52893814]I did not expect this to work at all...
[IMG]https://i.imgur.com/RhdZFI3.png[/IMG][/QUOTE]
Didn't even know C-Sharp allows you to fuck with pointers like that :v:
[QUOTE=Leystryku;52894163]Didn't even know C-Sharp allows you to fuck with pointers like that :v:[/QUOTE]
Look into the unsafe switch and be wowzered.
[url]https://docs.microsoft.com/en-us/dotnet/csharp/language-reference/keywords/unsafe[/url]
:snip:
Been messing with the UE4 particle editor a lot lately--it's pretty amazing. I'm not really good at art but I whipped up a little "main menu" for an RPG I'm working on.
[video=youtube;nGbK9STYk84]https://www.youtube.com/watch?v=nGbK9STYk84[/video]
Hopefully I'll have some actual gameplay stuff to show soon, too.
Some guy forked my ECS repo and has made a commit, which is really cool. Unfortunately it's before I've finished my update to C++17. So now I have to work on updating to C++17.
Working on a game based on real rocket factories like spacex, using unity and some placeholder assets
[video=youtube;60W4qcKOwh4]https://www.youtube.com/watch?v=60W4qcKOwh4[/video]
[QUOTE=Tleandrix;52900053]Working on a game based on real rocket factories like spacex, using unity and some placeholder assets
[video=youtube;60W4qcKOwh4]https://www.youtube.com/watch?v=60W4qcKOwh4[/video][/QUOTE]
You better finish this.
If you liked that image here's a falcon 9 made from rocket parts
[IMG]https://i.imgur.com/dO1CKY0.png[/IMG]
I'm still finishing some important mechanics for the game, currently working on workshops where engineers assembly multiple engines
I made a software 3D renderer from scratch for a class project.
[VID]https://my.mixtape.moe/xtmpmm.mp4[/VID]
I spent a lot of time trying to get my triangle renderer to draw edges just right only for it to be all fucked up by z-buffer inaccuracies. :s:
Working on a GMOD save for a YouTube video. POSTAL (game) centered, I kinda make a mock out of a certain other game...
[sp]I just hope my editor will render my video for once...[/sp]
@above Nice, reminds me of the PS1 (all the pixels and what)
[QUOTE=Stents*;52901379]I made a software 3D renderer from scratch for a class project.
[VID]https://my.mixtape.moe/xtmpmm.mp4[/VID]
I spent a lot of time trying to get my triangle renderer to draw edges just right only for it to be all fucked up by z-buffer inaccuracies. :s:[/QUOTE]
How is the performance, and how precise is the z-buffer?
Isometric is proving to be a huge pain in the ass. I wanted the front wall to be semi-transparent, but only after the fact did I realize they would overlap each other.
[img]https://i.imgur.com/mNW6s6h.png[/img]
Have a different wall sprite for the ones in the middle.
[QUOTE=reevezy67;52908065]Isometric is proving to be a huge pain in the ass. I wanted the front wall to be semi-transparent, but only after the fact did I realize they would overlap each other.
[IMG]https://i.imgur.com/mNW6s6h.png[/IMG][/QUOTE]
Ah, what's happening here is something called alpha blending.
Don't know what you are making it in but some shader fuckery is required.
Here's my go-to for Unity.
[URL]https://docs.unity3d.com/Manual/SL-Blend.html[/URL]
Using a shader sounds all well and good but layla is a fucking genius.
keep it simple sexy
the C++ standard sometimes does really dumb things: like how std::stod or std::stoi work. They only take string references, nothing else, and if fed a string_view it will perform a copy of the subrange (the const char*) you provided, stuffing it into a fresh new std::string. despite the fact that I don't perform a single copy, 95%+ of the inclusive samples in my last profiling run were lost to memcpy being used to copy new strings every single time I called a function to convert a string segment into a number (4 times per one of the 35,000 lines read). I don't even read line-by-line using the conventional std::getline- I manually advance the beginning of the single string_view object I create in the loop. So I see why things got slow now. Replacing the C++ functions with the old C functions means my whole file imports in a few seconds at worse, versus the 1min+ (even in release) it previously took :why:
What frustrates me is that there's no overload implemented for string_view, when it would be [I]so[/I] trivial to add (as far as I can tell, to be fair) and would encourage people to actually use string_view (which works great, when it works)
I got bored during class so I made a weighted random number algorithm.
You can adjust the amount of "sharpness" in the bias and even set the bias to cover a range instead of a single value.
These distributions are very satisfying to look at.
[img]https://i.imgur.com/gfmlsoW.png[/img]
Speaking of Layla...
Whatever happened to your "Source Engine Clone" thing you were working on with the WYSIWYG level editor?
[QUOTE=Lerlth;52919368]Speaking of Layla...
Whatever happened to your "Source Engine Clone" thing you were working on with the WYSIWYG level editor?[/QUOTE]
[url]https://facepunch.com/showthread.php?t=1577602[/url]
[url]https://sandbox.facepunch.com/blog/[/url]
I've been working on a little Arcade game where the main goal is to get as high as possible in a procedurally generated tower. I'm trying to see how much I can push out of Game Maker: Studio's terrible 3D capabilities, and so far I'm very satisfied with my results. So far I've managed 3D collisions, dynamic lighting, and I've also managed to get fully animated 3D models exported from blender working in the game. Ignore the sounds/models/textures used as they're all placeholder and they were grabbed off files I had lying around in my computer.
[video]https://youtu.be/l8asKmQW7l0[/video]
Also because the entire game takes place in a circular tower, pretty much half of my code is working with sin/cos. :hammered:
[QUOTE=buu342;52921386]I've been working on a little Arcade game where the main goal is to get as high as possible in a procedurally generated tower. I'm trying to see how much I can push out of Game Maker: Studio's terrible 3D capabilities, and so far I'm very satisfied with my results. So far I've managed 3D collisions, dynamic lighting, and I've also managed to get fully animated 3D models exported from blender working in the game. Ignore the sounds/models/textures used as they're all placeholder and they were grabbed off files I had lying around in my computer.
[video]https://youtu.be/l8asKmQW7l0[/video]
Also because the entire game takes place in a circular tower, pretty much half of my code is working with sin/cos. :hammered:[/QUOTE]
It's like an inverted Nebulus!
I kinda knew why my old CUDA library thing (the libnoise clone) was breaking, but I didn't realize the depth of my mistakes. Like this gem:
[cpp]
cudaAssert(cudaMallocManaged(&norm, dims.first * dims.second * sizeof(float)));
[/cpp]
followed by a device synchronization call wrapped in an assert macro, too. The problem is that my assert macro looks like this in anything but debug mode:
[cpp]
#define cudaAssert(expression) ((void)(0))
[/cpp]
gg nore, past me. No wonder [I]everything[/I] broke outside of debug mode. [U]Every[/U] single call to cudaDeviceSynchronize was wrapped in this assert macro, in my infinite genius. And a few calls to cudaMallocManaged. Also, I didn't free the memory in that first snippet. Which explains why my memory use always shot up, since this was used a ton to normalize data between several key stages in my noisegen pipeline
the synchronization was admittedly a bit weird, as it's not really required for my simple kernels - just for the more complex kernels (sometimes it'd break, sometimes not). since this wasn't being done, whenever I tried to get the data pointer from one stage to feed to the next I got a pointer to incomplete data (or CUDA threw an exception, since I was accessing data currently in-use on the device). this had me stuck for at least a good week, and it was such a silly+dumb mistake to make lol
[thumb]http://carp.tk/$/pCRdBp.png[/thumb]
SVG to G Code works well, alright?
[QUOTE=cartman300;52925269][thumb]http://carp.tk/$/pCRdBp.png[/thumb]
SVG to G Code works well, alright?[/QUOTE]
Dickbutt when?
Cloth sim for uni assignment:
[video]https://youtu.be/jXl-L53PzZU[/video]
The integrator was already provided and basically just does explicit euler so it's a bit unstable and still thinking about whether or not I fucked up damping but it seems to be working as well as intended for the assignment. Assignments for this class are super rigid in terms of where we add our code and what state we can access and modify so there's pretty much no further room for any improvements there. Haven't done cloth physics before so kinda surprised with how simple the code is for an okay-ish result.
[QUOTE=cartman300;52925269][thumb]http://carp.tk/$/pCRdBp.png[/thumb]
SVG to G Code works well, alright?[/QUOTE]
Inkscape has a plugin to do that
[QUOTE=DrDevil;52926191]Inkscape has a plugin to do that[/QUOTE]
Mmmmm yep, and it throws exceptions for me. This is just a part of a bigger project that gives you a nice API to convert text shapes, heightmaps, SVGs and stuff to a GraphicsPath which in turn gets converted to G Code.
[thumb]http://carp.tk/$/MpdWbH.png[/thumb]
I've made a number of large changes to PortalKit in anticipation of the ongoing sale [URL="https://www.assetstore.unity3d.com/#!/content/81638"](check it out!)[/URL],
but the largest one is a total UI rework, VRTK re-working, and recursion support with a new rendering method that causes extremely low vram overhead (especially compared to older versions). Extremely mad props to [URL="https://twitter.com/madgvox"]@Madgvox[/URL] for doing a ton of the UI and guiding me through the rest.
[IMG]https://i.imgur.com/NA70URQ.png[/IMG]
Also, the feedback system uses a super hacky last-minute IRC-based solution since our server contracts had issues a couple days before this update was supposed to roll-out and that was the only even remotely reliable solution.
[IMG]https://i.imgur.com/jTGuaKu.png[/IMG]
Sorry, you need to Log In to post a reply to this thread.