Learn OpenGL, extensive tutorial resource for learning Modern OpenGL(learnopengl.com) |
Learn OpenGL, extensive tutorial resource for learning Modern OpenGL(learnopengl.com) |
After you've learned it, you can start learning CUDA if you want to do some more low-level compute stuff on the GPU (sorry, but you should just buy an NVIDIA card, CUDA is just that good). Or if you actually want to just make things but want to use a nicer cross-platform graphics API than OpenGL, then I recommend SDL3. (Or use Metal if you want to make macOS exclusive apps - it's actually a quite nice API)
Vulkan or DX12 are currently flawed APIs that are unnecessarily complex and doesn't even match the performance characteristics of current-gen hardware anymore. (see the titular post: https://www.sebastianaaltonen.com/blog/no-graphics-api). However if you want a computer graphics career (either in the game industry or in other niche domains) having experience with these APIs will be beneficial since these are what many production apps are currently stuck with - though honestly the job market for graphics really suck nowadays. (The biggest sector was triple-A game companies with their own engines, but the game industry is imploding right now...)
[0] https://github.com/floooh/sokol [1] https://wiki.libsdl.org/SDL3/CategoryGPU
Then I realized you're just writing code that executes on all pixels sequentially. That made it easier to understand
Opengl meanwhile is significantly easier, and will give them the important terminology and math required to do graphics. Once the student has learned opengl and wrote a small renderer with a few basic techniques, moving to vulkan or DX12 will happen far more smoothly.
In particular, learnopengl explains a lot of basics like transformation coordinates, what a mesh is, and other similar "basic knowledge", while all vulkan and dx12 tutorials skip through that because they are meant for a much more experienced audience.
You want to start with OpenGL because the API is organized roughly around these high level concepts. Starting with something like Vulkan means you're dropped straight into hundreds of lines of low level boilerplate setting up VkCommandBuffers and VkRenderPasses and VkPipelines and shaders, and all this GPU programming, and it's all just bewildering if you're used to thinking about things at the high level first, or just want to draw a fucking triangle on the screen.
I would agree if by modern you mean DX12 or Vulkan, but WebGPU (and maybe even Metal) are quite decent.
Even if OpenGL doesn't get new features what exists now will continue to work for decades.
The OpenGL API with its binding-based approach is also very confusing for beginners, but the replacement Direct State Access API came too late (after Apple already decided they would deprecate OpenGL, so it never got supported on macOS).
And the fact that the GLSL parser and compiler are part of the driver means there are tons of hardware-specific bugs and miscompilations. Intel integrated drivers on Windows are notorious for being especially buggy.
All of those make OpenGL a pretty poor target for learning. Personally, I would recommend WebGPU to people who want to get into graphics programming. It's very similar to DirectX/Metal, and like a more streamlined Vulkan. The WGPU implementation of WebGPU has a C API, which also has C++ headers, and a native Rust API. And you can also use WebGPU in your browser from JavaScript, if you don't know any native languages.
Another (related) problem is that the many sediment layers that have accumulated over nearly three decades are not clearly separated (and Vulkan is starting to suffer from the same problem btw, there's always at least five different ways to do the same thing, three of which are outdated or not recommended to be used on specific GPU architectures, and the other two have complicated relationships and interdependencies with other redundant or optional features, it's apparently some sort of Khronos curse to always create the biggest possible mess when it comes to 3D APIs).
On Windows (maybe even on Linux via Proton), starting with D3D11 makes a lot more sense, or on macOS with Metal v1. Both APIs are 'close enough' to modern 3D APIs to carry a lot of the knowledge over, but without being too low-level like Vulkan or D3D12, WebGPU running in browsers is also a good choice (even though it inherits some bad design decisions from Vulkan 1.0 when it comes to resource bindings, and is in some places actually slower than WebGL2).
If you are going to use Vulkan, check out https://howtovulkan.com/ Vulkan started with a lot of compromises to make mobile hardware happy at the expense of making everything overly complicated on desktop. Over the past decade, desktop devs have managed to get a lot of features added to make Vulkan on desktop more sane. How To Vulkan covers that newer approach. This recent video "It's Not About the API" https://www.youtube.com/watch?v=7bSzp-QildA shows how simple it can be if you let it.
And, if you are on a Mac, folks who use Metal like it a lot. Don't worry about lock-in. Once you learn the basics, knowledge is easily transferable to DX12 and Vulkan. You should plan to write 3 or 4 renderers to throw away anyway :P
For me, a much stronger argument against OpenGL is that is requires a global state. This often leads to bad design, is miserable to multi-thread, and is rather tedious to port to Vulkan. I have spent several years fighting with multi-threading applications doing all sort of things in OpenGL and it is not great. Yes, Vulkan is extremely painful to set-up and you have to think about a million things you might not really care about directly, but honestly, I would also recommend against starting anything serious in OpenGL. To get your hands wet maybe, as with a couple lines you can get something running..
But I can agree with CUDA. CUDA is great, in a large part because it actually offers an easy to use API.
It's also much easier to grasp than Vulkan, Metal, or Direct3D.
I wonder if somebody has written a adapter layer to write old OpenGL and translate it on the fly to Vulkan?
Apple themselves also don't ship a native OpenGL driver anymore, only a layer that translates OpenGL to Metal.
Even if you use OpenGL you still need to write shaders.
Apple's OpenGL on macOS and iOS has been a shim over Metal for a very long time. The feature set is stuck somewhere between GL 3.x and 4.1 though (e.g. no compute shaders), but not for technical reasons.
Vulkan and DX12 are currently flawed APIs that are unnecessarily complex and doesn't even match the performance characteristics of current-gen hardware (see the titular post: https://www.sebastianaaltonen.com/blog/no-graphics-api) - if you want to really learn the low-level details you should start diving into something like CUDA or get into graphics driver development (start by reading Mesa open source driver code - which you will then learn why Vulkan is flawed)
If you're looking for fundamental material, the last thing you want is learning a very (not slightly) outdated and weird API.
> Vulkan and DX12 are currently flawed APIs that are unnecessarily complex and doesn't even match the performance characteristics of current-gen hardware
And OpneGL is worse
it also will run just fine on windows or Linux using Proton with no issues
if I had to learn Vulkan from scratch instead, that would have been a whole different animal...
I think I once bookmarked an article which takes an interesting approach, namely doing all rendering with Vulkan compute! Does that ring a bell to anyone? I can't find it again. It seems very appealing to me. It may not be the best way to approach graphics, but I know a lot more about computations in general than I do graphics in particular. And knowing Vulkan Compute would help me in lots of ways, so it I can also use it for graphics that's a nice bonus.
The biggest issue with c# dev is OS specific window management. CLR doesn't have good wayland support yet if you're trying to do this from a modern Linux distro.
I definitely recommend it to anyone interested! Honestly one of the biggest things it helps with is explaining the setup boilerplate for OpenGL.
Unfortunately WebGPU is like 5 years behind OpenGL 4.6, which is already 8 years old. WebGPU is pretty ancient.
You need a native build for proper GPU debugging, or do shader printf debugging style.
However, it is what is available.
Is it the AZDO stuff? Quite interested, since there isn't really much information on the Internet about it rather than some slides and GDC videos.
Instead, the "modern" APIs allow you to leverage your pre-existing knowledge of "allocate arrays of structs and start indexing them." That's not trivial. But, it is already familiar. And, it ends up a lot better in the end compared to "Invoke whole lot of functions to manipulate a hidden state machine."
You can find a little info on Modern OpenGL at https://github.com/fendevel/Guide-to-Modern-OpenGL-Functions, https://juandiegomontoya.github.io/modern_opengl.html, https://ktstephano.github.io/, https://patrick-is.cool/posts/2025/on-vaos/
DX11 suffers from the problem that there aren't that much high quality material comparable to learnopengl.com - I really don't like RasterTek which just dumps code at you without explaining things properly. Same for Metal - the best way would be to just download and read the example source code from the official Apple site, but it's rather unfriendly for beginners.
The fundamental problem is Metal 3 is just too high-level to be able to emulate all of Vulkan's behavior. The new Metal 4 API (which is more low level and similar to Vulkan in many ways) might have improved things recently, but sadly MoltenVK hasn't been rewritten to this new API yet.
> Even if you use OpenGL you still need to write shaders.
A data point: SDL3-GPU maintainers announced that they will not give a shit about Android support, since there's just too many devices that haven't properly implemented the Vulkan spec. (https://github.com/libsdl-org/SDL/issues/12652#issuecomment-...)
Another data point: the current maintainer of the renderer portion of the Godot engine suffering through all the bug reports from Android devices (https://github.com/godotengine/godot/issues?q=is%3Aissue%20s...)
I've never seen a more detailed tutorial than LearnOpenGL that actually goes through the concepts of graphics programming thoroughly and actually make things other than just "learn how to use the API". That alone should be enough to cement the site's longevity.
> And OpneGL is worse
It's never late to learn Vulkan / DX12 after learning learnopengl.com! The value of learnopengl.com isn't in the APIs, it's about learning the basic concepts of graphics programming. If you start learning with Vulkan without any prerequisite knowledge, you'll be bogged in low-level details from the start that isn't really related with learning the actual fundamentals. (And you'll probably have to unlearn Vulkan and DX12 again once a new API has surfaced)
Why would you do that? Why would you want to learn an API for graphics cards from the 90s instead of an API for the graphics card from the 2010s?
> The value of learnopengl.com isn't in the APIs, it's about learning the basic concepts of graphics programming.
Then you should learn concepts of graphics programming that isn't defined with code snippets like
#version 330 core
struct Material {
> And you'll probably have to unlearn Vulkan and DX12 again once a new API has surfaced
Strange. You "have to unlearn a newer API when an even newer one has surfaced, but you should learn this ancient API anyway".
Is it? All the fundamentals are tied to how it's done in OpenGL with somewhat sparse explanations in between. And many "fundamentals" are literally something like
#version 330 core
struct Material {https://learnopengl.com/PBR/Lighting
Even the code snippets start with "#version 330", if you implement these in another tech stack (say vulkan+slang) the relevant shader code would look very similar.
I disagree. OpenGL is truly bad, but there is no API in this world as bad as Vulkan.