Hacker News Reader

Hacker News

Comment by gdxhyrd | Hacker News Reader

Parent Full thread

gdxhyrd

No, OpenGL is not "on the way out". Please back up your claims.

View on HN

“Apple deprecates OpenGL across all OSes”: https://www.anandtech.com/show/12894/apple-deprecates-opengl...

Apple is not Khronos.

They haven't supported OpenGL for years anyway, so the fact they deprecate it now is irrelevant.

A cross platform library loses its potency if it doesn't work cross platform.

OpenGL does not stop being cross-platform just because it does not work natively in 1 platform.

In any case, abstraction layers for macOS already exist.

And, going by your definition, if Apple only allows Metal, then no API will be cross-platform ever anyway.

OpenGL was never supported on games consoles, so....

The GL tools rot is already starting. E.g. the Radeon GPU Profiler supports Vulkan, D3D12, OpenCL, but not OpenGL:

https://gpuopen.com/gaming-product/radeon-gpu-profiler-rgp/

That profiler is meant for low-level debugging as its own description says. That is why it does not support DX11 either.

AMD's OpenGL tools don't look too active either, e.g. the last release of CodeXL is from 2018, and this only updated dependencies or removed functionality, for instance this nugget from the release notes:

---

* Removal of components which have been replaced by new standalone tools:

* FrameAnalysis - use https://github.com/GPUOpen-Tools/Radeon-GPUProfiler

---

...that Radeon-GPUProfiler is that tool which has only D3D12 and Vulkan support.

Look around for AMD's OpenGL activity more recently, there's not much, which isn't surprising because they've been lagging behind NVIDIA with their GL drivers since forever. I bet they're eager to close that chapter.

NVIDIA seems more committed to GL still, but without having all GPU vendors on board to continue supporting OpenGL, Khronos won't be able to do much to keep it alive.

AMD has always lacked tools and has never produced much on the software side.

That is not news, and that has nothing to do with the state of OpenGL.

Please, stop spreading misinformation about OpenGL.

> but without having all GPU vendors on board to continue supporting OpenGL, Khronos won't be able to do much to keep it alive

Please don't make things up. OpenGL 4.6 was released in July 2017. According to Wikipedia, modern AMD and NVIDIA cards both gained driver support for it in April 2018. Intel drivers have support since May 2019.

(https://en.wikipedia.org/wiki/OpenGL#OpenGL_4.6)

OpenGL absolutely is a legacy API today. It has been an awful impedance mismatch to modern GPUs for about a decade now.

It's a high-level API. It went through N generations of graphics hardware for 30 years (40 if you consider IRIS GL).

And it became increasingly less of a good fit with each one. By now, it is ridiculously mismatched to the task it needs to perform.

Just leave it to die.

OpenGL is certainly used by a number of creative apps like Photoshop, but you can't deny that the number of games released with OpenGL is down significantly since say 2010.

A lot of games and apps are being released using OpenGL today.

Then there is WebGL and OpenGL ES, widely used everywhere too.

So, no, it is not going anywhere. In fact, Khronos themselves have said so.

I didn't say that there were no OpenGL games being released in 2019/2020 but that the ratio is definitely skewing away from OpenGL. "A lot of games" seems like a stretch compared with what the numbers used to be.

Also I think it's obvious but I'll say it anyway; it's not up to khronos it's up to developers. It doesn't matter if they continue to support OpenGL if developers move to some combination of dx11/12, Vulkan, and Metal.

WebGL and OpenGL ES aren't OpenGL. Their uptake or lack thereof in other arenas is orthogonal to whether people are moving away from OpenGL for game development.

Considering UE4 and Unity both support OpenGL, yes, a lot if not the majority of games.

WebGL and OpenGL ES are pretty much OpenGL. Formally they are different standards, but they are based all in OpenGL: if you know one, you know the others pretty well, which is what is important for developers as you agree in the second paragraph.

Once again, that's support not implementation. Developers could use OpenGL in Unreal. They could, but they haven't in general. Vulkan is the default, and is the more common choice.

If it were true that lots of games were picking OpenGL, why is it so hard to make a long list of them? It's easy to make such a list for vulkan.

Reply on news.ycombinator.com