/agdg/ - Amateur Game Development General

Where's that link that explains what happened to the agdg community after the exodus? About splitting into groups and whatnot?
I can't find it.

tiles.png
[Hide] (122.9KB, 960x540)

>>2259
I have slowly slogged my way to rendering tiles again. Drawing what's on this picture takes about 15 milliseconds (generating tile layers is 7-8 ms, loading the layers into GPU buffers and drawing them is another 7-8 ms). I'm looping through every tile on the visible chunks and generating sprites for them every frame, which I'm obviously not supposed to do, but I just can't help the feeling that this should be much faster, and that I'm still missing some extremely fundamental secret that nobody is telling me.

I haven't tried >>2263 yet, getting this to perform well is more important.

1637374506247.png
[Hide] (59KB, 259x195)

>>2343
Yep this instantly killed all of my motivation to work on videogames again. Obviously I can pre-build the buffers and draw them to make this way faster, but you can't fucking tell me that computers aren't fast enough to properly render 30k dynamic objects even when they never have to change shaders or textures. What am I doing wrong? I'm literally just writing the instances into an instance buffer and telling opengl to draw them. I don't fucking get it.

>>2344
We'd have to look at your implementation to find the problem.

>>2344
I was too tired and frustrated to even try making an optimized build, but I just did and didn't realize it would have such a big effect: generating tile layers takes 1ms, and rendering them takes 3ms.

Anyway, I'll have to redo parts of my rendering system for the 500th time. One of the "simplifications" I've made is that I'm calling functions that create everything for VAOs from scratch. However what I want to do now is to create a separate buffer for each chunk, but I only need the instance buffer to be different, all of them should use the same vertex and index buffers. Another day of me wishing I could just do gpu_alloc() and point my draw calls to it instead of initializing buffer objects and attributes and linking memory buffers to data structures.

>>2345
The implementation is in like 870 separate places, I doubt there's anything useful I could post.

>>2346
>I was too tired and frustrated to even try making an optimized build
Anon...

So I don't exactly remember how long ago I made Kasha. I'm pretty sure I've had her since at least 2001 but maybe longer? Or not? She existed in some form since I was in high school and probably earlier. 

Kasha's a catgirl with the power to shapeshift her tail into anything she wants. She usually does this by drinking a milkshake and different milkshake flavors have different effects. I planned to use this as the main gimmick of her platforming adventure but it would also make the minigolf game fun as well.

Kasha lives with Milchi in Purraya Province, the western coastal province of Tabbyoko. Together they live in Mellow Rock which is a huge fancy estate built into a mountain sculpted in the shape of a tiger. Kasha enjoys milkshakes, watching cartoons and any game that involves balls. Milchi is the princess of a tribe of tall cat people named Laimonans whose parents were entrusted with Kasha's care when she was a baby. Milchi has the kindest heart of anyone so naturally she fell in love with Kasha and took the responsibility of taking care of her entirely upon herself, and now they live together in Mellow Rock while her parents continue their royal tasks back home on her island.

Occasionally they are menaced by Dai, a big fat snake lady from Milchi's childhood who is the princess of an opposing tribe of snake people called Sunbaskers, and constantly harasses them through selfish mischief. Dai tends to be a recurring antagonist in their lives.

>>2284
>high school
Real chuuni hours
I liked the catcher game, too bad mods are fags

There is no "mature content" in this post or anywhere in any of the things I posted. All of my images were safe for work.

I can see the mods here are just as gigantic faggots as anywhere else, I can see I'm not welcome here and I guess I have nowhere else to post then. That's perfectly fine. Whoever mods this place can eat shit and die.

Posting your disgusting fetish and begging for gibs to make more does make me quite uncomfortable.

>get kicked out of every place you post
>the problem is surely with them, not me!

banner.webm
[Hide] (1018.5KB, 1000x600, 00:20)

>>2070
>>2073
Released 1.0 of this just now.
The reason I experimented with vectors >>2219 is so I could draw the favorite stars.

The only real thing missing is performance, but I can't be bothered to do anything since this is already fast enough for all of my needs. The amount of effort required to manage a path/file cache wouldn't be worth what I'd gain.

>>2335
Neat! Glad you've made such progress with it. Any plans for new projects now?

>>2336
I jump around projects randomly so I can't predict anything. Maybe an imageboard.

---

Speaking of making an imageboard, I'm thinking of how to store data in an efficient way in a database/whatever. A lot of strings are very short, so it feels dumb to store a 4-byte or 8-byte integer for their length, giving them a fixed length like 200 is also wasteful. On the other hand, if you use a small integer, then you may run into limitations. So, to make things easier and get all the advantages, I came up with a way to read and write variable-length integers, inspired by UTF-8. I can just use this and not have to think about it:
#define swap_endian_16(x) __builtin_bswap16(x)
#define swap_endian_32(x) __builtin_bswap32(x)
#define swap_endian_64(x) __builtin_bswap64(x)

static u64 flexint_read_u64 (void* data, u64* value) {
	if (*(u8*)data < 0b10000000) {
		*value = *(u8*)data;
		return 1;
	}

>>2337
Neat! I personally prefer the horrors of compile-time dispatch using Sepples templates, but this looks efficient.

>>2337
One downside of this is that it might be harder to resize data since the beginning of the string may need to move, you have to avoid overwriting the memory from the beginning of the string and such. Honestly, I've had variable length ints my mind for a long time but I can no longer remember what exactly I wanted them for. It was probably for some kind of string lists, arrays of arrays, maybe even data streams for an MMO, or to compete with JSON which technically only needs 3 bytes/characters to separate strings ("",).

>>2289
no you're just autistic

tsumugin.jpg
[Hide] (66.5KB, 1080x969)

>>2288
the most normal thing in the world is to call someone a retard when someone does something retarded

normal.png
[Hide] (33KB, 453x306)

>>2296
>the most normal thing in the world
What about normal vectors huh? Didn't think about those, did you?

>>2298
Explain normalmaps in one paragraph or less.

>>2298
You got me, those are more normal.

>>2305
Easy: Mercator projection.

Heads up to moderation: the guy who tanked the /agdg/ thread over on zzchan apparently thinks he's gonna try his shit here too now. His schtick is to post gifs and shit of a bunch of squares moving on a screen and then some inane non-question like "what do I do now? I don't get the maths" or whatever. Keep an eye out and delete/ban on sight please, thanks.

>>2238
This is him is you don't recognize the posting style: 
>>2236
>>2237

>>2238
>>2239
Thanks for the headsup, Anon. Yes, he's been abusing Trashchan for a few days now.

>>2238
I still find it baffling they refused to do anything about it.

>>2244
It's appreciated.

>>2247
>I still find it baffling they refused to do anything about it.
I remember seeing that guy in the zzz/v/ /agdg/ threads and his posts were promptly deleted almost every time they appeared. Did I miss something?
I figured those threads slowed down due to the autist posting his DA-tier scribbles of uncanny women.

>>2105
Newer doesn't always mean better. There's people who still make amazing stuff with just C or "C-like C++" (AKA C with only a select few C++ features) because the fundamentals of programming haven't really changed in decades, it's much more important to nail that part than some C++ features.

>>2108
I haven't read the books but I would avoid the Java ones because that's a pretty abandoned language these days and was always a mistake in a lot of ways. It may be useful if you want to get a job maintaining some old ass crapware that was built on Java though.

Visual Basic stuff will probably get you locked into Microsoft's ecosystem, which is extra bad since Microsoft has been heading straight into the trash in recent years. Some of that stuff is probably obsolete too since Microsoft keeps updating their shit.

>>2109
>Newer doesn't always mean better.
No, but C++11 got some new features to the point it might be considered as a different language (lambdas, vararg templates, range-based for, constexpr functions, etc. Concepts can also be useful if you do templates, but they're a newer additions and assumes you're not stupid. Of course still no static reflection, so you'll be forced to roll your own for anything but the most trivial programs, but that's just how C++ evolves, useful features have the lowest priority)
>fundamentals of programming haven't really changed in decades
Yeah, but these books look more like teach language X instead of teach programming fundamentals.
>I would avoid the Java ones because that's a pretty abandoned language
Thank you for contradicting yourself in one post.

>>2113
I'm not against Java because "it won't get new features", I'm against it because there's no future for it. There will be less and less jobs for it, less of a community and discussion and help, and the virtual machine that runs Java will probably have less support in the future, who knows if it runs on future consoles or phones.

>>2114
>no future for it
I wouldn't say it so clearly. Sure, not as popular as it used to be, but lot of enterprise garbage are still written in java. Visual Basic is much worse in this regard (not that I'd miss any of them, both languages are an abomination).

>>2266
You can only get a lightweight engine or a drag/drop editor.

>>1787
Great stuff.

Parallelizing_the_physics_solver_–_BSC_2025_[Kvsvd67XUKw].webm
[Hide] (10.2MB, 1280x720, 00:47)

The developer of Teardown did a talk about their new physics engine.
https://www.youtube.com/watch?v=Kvsvd67XUKw

657e7db79fbfdb31b986679fcccd308bc09565c5bbf44f969bc87c7adda33564.jpg
[Hide] (62KB, 564x635)

I've come to a very enlightening conclusion on gamedev and it's really improved my outlook on how successful my game will be. It's based on the idea that most people (90% or more) involved in gamedev are idiots. These are the people that are permanently stuck in tutorial hell/decision paralysis instead of actually making something and getting hands-on experience, trust fund kiddos that do nothing but throw money at third-worlders on Fiverr and Upwork and say "make my game for me" then get really mad when none of it looks good or plays well, people that don't understand why their totally unique match-3 puzzle game with pastel colors only sold two copies, one of which was refunded, or people that actually have a good game but only decided to start marketing weeks before release with some halfassed reddit posts and nobody knows it exists.
So when you read articles like "Most games on Steam never make a profit" it seems demoralizing at first, but when you look a bit deeper you realize that most games released on Steam are asset flips, AI garbage, puzzle games, or games that otherwise had zero effort put into their development or marketing. The few remaining games are ones that actually had effort and thought put into them and are doing pretty well.
With all this in mind, if your game is:
>unique
>fun to play
>looks good
>has a modest amount of mar

>>2222
The vast majority of people are retarded. There's nothing surprising about that, but it's hard to internalize just how far-reaching consequences that has, other people's retardation can warp your perception of reality. Most people really have no idea what they're doing or talking about, and lack even the most basic ability to interpret the world around them or their own actions and their consequences.

As long as you care about making a good game, succeed in making it, and don't completely fuck up the content that is visible on your store page, it will probably find success automatically thanks to Steam's algorithm and word of mouth. You only "need" marketing if your game is ok but not great, or if your game is garbage that relies on meme streamer appeal or virality to make anyone care about it.

>>2222
This has also been a revelatory experience for me, in many things other than game development. With how many generation-defining indie games have released in the last few years (I consider 2025 to be 2 AP; 2 years After Pizza Tower) it can seem daunting to ever imaging having the same level of cultural or commercial success.

But I'm finally trying to get around to doing something with my free time and ideas. If the number of stupid people who've stumbled onto success is any indicator then it can't be that hard.

ClipboardImage.png
[Hide] (72.6KB, 408x206)

>>1587
Are you trying to invent cataclysm:dark days ahead?

>backpacks
Why not just make everything into a storage? All existing items have to be stored somewhere, and a lot of items are both storages and "items". So simply register every item, by ID/pointer into its storage space. On loading, recursively load all used storages, starting from "the world". 
Its actually important part of the interface. Depending on number of items/buttons, game might start lagging, if every button pressed checks every other button location against it. You will need to separate them into containers anyway. 
And with everything being a container of their own, you can have "hidden" part, where item stores materials its made of.  

>how to ID
You either use half assed system, or make a proper all encompassing item ID system. 
> I want to refer to the inventory by it's name
You dont mean, you literally search all existing items via std::string or something like that?

items.png
[Hide] (34KB, 752x483)

>>1587
>>1588
>create_item()
>create_inventory() 
I just realized that the way I explained it doesn't make any sense because I left out context. Those functions in reality are create_iteminfo() and create_inventoryinfo().

The way my stuff is organized is that I have objects and infos. The info represents the object's type and has all the static information about it, like name and description and sprite, you only modify those at game startup. The object has per-object information, like how many items are in an item stack, they are actual objects that exist in the world.

Pic related are the actual, un-edited data structures for my items. What I'm talking about is creating an Iteminfo, and connecting the relevant Inventoryinfoid into it. I don't need or want to store the inventory info's string ID into the item info, I just want the numerical ID, but the item info may be created before the inventory info so the numerical ID wouldn't exist yet.

>Are you trying to invent cataclysm:dark days ahead?
Not sure what makes you think that.

Test.

c0f9b45354cd49fbcb8f0395a73230637a76596de76df37bb70041c11c947e7c.png
[Hide] (60KB, 1200x1000)

New bread.
>>1605
>>1605
>>1605

>>1587
You need a better vision of what your game should be about, what is it called?

>>1896
Possible but probably not? You sound like you use giant cpu towers and high-end gpus which are giant anyway.

proxy-image.png
[Hide] (22.3KB, 3840x2160)

What one is better (for game development), the RX 7700 XT or RX 6750 XT?

I'm planning to pair it with an Ryzen 7 7700 and 32gb DDR5 6000mhz.

The price difference between these two cards is about 63 euro.

I don't think anyone on any imageboard has the required level of hardware and software/engine performance knowledge to be able to answer this.

>>1953
Should be fine either way. It only matters for working with blender, and in a minor way. Most new hardware is more than good enough for game dev, except when its something randomly unsupported.

>>1953
Both are fine. You usually just want something with enough VRAM, low power draw and a fast enough core.
However, the 7000 series/RDNA3 has a flaw with rendering proper 1080p with av1, I believe, and AMD just gave up on fixing it since it was a hardware/architectural issue.