6 ms·
About Godot4, Vulkan, GLES3 and GLES2
- lazypenguin 5y agoAs a non-contributor but past user of Godot, the different rewrites of the renderer is what turned me off from committing to the engine long term. I have no issue with Godot staying modern (I.E adopting Vulkan) but my perspective is that this decision was made to maximize graphic fidelity rather than to maximize usability of the engine for the users. For me the biggest appeal of Godot was it was approachable and usable. I don’t care if it can compete with unreal or unity on a visual level because there is more work needed in other areas of the engine first. Perhaps the devs felt that the rest of Godot was “good enough” and the renderer was a higher priority but I did not feel that was the case. In the end this was a decision I did not agree with but I am sincerely happy they are making good progress and I’m looking forward to trying out Godot 4.0 once it hits alpha!
- lasagnaphil 5y agoI think graphical fidelity can also be part of usability. Artists always seem to complain about things like waiting for baking lights to finish, shadows having weird artifacts, lighting not appearing naturally and have to resort to using hacks, performance problems preventing the use of certain assets and effects, etc. An overall better renderer would remove lot of the pain from tweaking your art to fit the quirks of the engine. For example, things like SDF-based global illumination would be a godsend to those making outdoor scenes.
- enos_feedler 5y agoRefreshing to see that the GPU stack might be converging on GLES3+ at a high level and Vulkan at the low-level. Of course Apple is left out of this, but getting some translation over to Metal is the best we can do.
- andrewmcwatters 5y agoI find their prioritization of features also very weird. I can't start an empty project and get multiplayer out of the box, but my team's 7 year old (?) 2D game engine[1] can do this and have client-side prediction set up for you before you even insert your own game code. We're months away from shipping a next major release, and for as many collaborators work on Godot, they've basically been years behind any of the features I or anyone on my team actually cared about. We have lag compensation that works over 3G, but if you asked anyone in the Godot community how you'd pull that off, it'd be some random guy not associated with the project, on their forums, with some code snippet he shared with you that is a far cry from first-class support. Why focus so much on rendering functionality? If you have a team that's actually remotely capable of generating assets at the labor velocity you would need to actually release a game, that team would almost be guaranteed to use Unreal anyway. It's really incredible how much game software lags in progress because of where development teams choose to put their efforts. [1]: https://www.planimeter.org/grid-sdk/ https://www.planimeter.org/grid-sdk/
- nextaccountic 5y agoTruth is, as an open source project, Godot would need people interested in multiplayer to implement that stuff. :(
- okamiueru 5y agoIs your sentiment not a little unreasonable? You seem to disagree with the godot maintainers on where the effort should be placed, but, this is as open source a project as they get. Improving one aspect doesn't belittle all others, nor does it really hinder it. As unsurprising as it might be, they are improving godot in the ways that they can, because they want it to be better. This is unrelated to the existence of UE. I'm sure the godot community wouldn't mind if you shared your experience on game networking, and possibly contributed in that regard. I myself would appreciate it, as with you, I've found it to be limited. The underlying components and language features are there, but how to tie it together with redundancy and prediction, and a bunch of other tricks of the trade, is not trivial.
- andrewmcwatters 5y agoNope, this team makes 5 figures in donations a month, and they actually spent time writing their own programming language. Total waste of labor. 100%. Let's put it this way, say I gave you $250,000 and said in a year I want a game engine. What would you spend your labor on? It would not be writing a programming language.
- djmips 5y agoWhy the downvotes. I feel like there is some truth here. The world doesn't need another custom game scripting language.
- lasagnaphil 5y agoThis has been repeatedly said throughout Godot’s development history, but there are several crucial reasons as why the team was motivated to create their own scripting language. From the FAQ (https://docs.godotengine.org/en/stable/about/faq.html#what-were-the-motivations-behind-creating-gdscript https://docs.godotengine.org/en/stable/about/faq.html#what-w...): In the early days, the engine used the Lua scripting language. Lua is fast, but creating bindings to an object oriented system (by using fallbacks) was complex and slow and took an enormous amount of code. After some experiments with Python, it also proved difficult to embed. The main reasons for creating a custom scripting language for Godot were: - Poor threading support in most script VMs, and Godot uses threads (Lua, Python, Squirrel, JavaScript, ActionScript, etc.). - Poor class-extending support in most script VMs, and adapting to the way Godot works is highly inefficient (Lua, Python, JavaScript). - Many existing languages have horrible interfaces for binding to C++, resulting in large amount of code, bugs, bottlenecks, and general inefficiency (Lua, Python, Squirrel, JavaScript, etc.) We wanted to focus on a great engine, not a great amount of integrations. - No native vector types (vector3, matrix4, etc.), resulting in highly reduced performance when using custom types (Lua, Python, Squirrel, JavaScript, ActionScript, etc.). - Garbage collector results in stalls or unnecessarily large memory usage (Lua, Python, JavaScript, ActionScript, etc.). - Difficulty to integrate with the code editor for providing code completion, live editing, etc. (all of them). This is well-supported by GDScript.
- modeless 5y agoMaking GLES3 the baseline and dropping GLES2 is the right thing. I'm glad that Godot was able to make that decision now that Safari is getting WebGL 2 (GLES3) support. I spent a lot of time on the implementation of WebGL 2 in WebKit and it's good to see it helping people.
- joeyjojo 5y agoIs Safari actually getting WebGL2 support? I looked into this around 6 months ago, and what I gathered was the WebGL2 support was based on rumours from a Twitter post, and the supposed checkbox to enable WebGL2 on iOS actually does not enable it. I still can't find any solid info about WebGL2 coming?
- Jasper_ 5y agoApple has switched their renderer to using ANGLE, and the iOS 15 Beta shows it in action. I've even reported bugs to the WebKit team and see them fixing it.
- modeless 5y agoYes! The announcement was a bit buried but WebGL 2 was officially confirmed by Apple at WWDC 2021 for Safari 15 "on all Apple devices": https://developer.apple.com/videos/play/wwdc2021/10030/?time=1167 https://developer.apple.com/videos/play/wwdc2021/10030/?time... Test it now in the iOS betas or in Safari Technology Preview. Note that there is a bug in the current iOS beta (which I expect will be fixed before release) that prevents WebGL 2 from being enabled on older A-series processors.
- yann2 5y agoUnrelated but having played with it recently what I really like about Godot is its 35MB or something to download and start playing with. Meanwhile every other engine is 7 zillion petabytes or whatever. For getting kids into programming, its a beautiful thing.
- hn8788 5y agoThere's a reason for the large size of other engines though (they're still only ~10GB if you don't plan to release on every platform). You have to implement pretty much everything yourself in Godot, so if you're going to do anything non-trivial, there's going to be a lot of work to implement things that are already included in the other engines. There was a post on the Godot subreddit recently by a team making a 3D ARPG, and they said that if they had to do things over again, they'd use Unreal Engine due to the amount of work and engine customisation they've had to do just to get standard features that are included in Unreal. Godot is definitely a good introduction to game dev, it's the first engine I used and it's very beginner friendly, but you quickly end up hitting roadblocks once you move past anything more complicated than a 2D platformer.
- TillE 5y agoIt's absolutely wild to compare Godot with UE4, they're just serving totally different purposes. If somebody chose the wrong option for their project, yeah that's a huge mistake. 2D games are hardly some sad niche for beginners though, it's an enormous fraction of non-AAA development, and there really aren't a ton of good 2D engines out there.
- TillE 5y agoIt also compiles from scratch pretty fast, and the code isn't hard to read. I was able to find and fix a minor bug I'd been experiencing within a couple hours. It's a beautiful middle ground between simple game frameworks and big clunky engines. It's precisely what I'd been looking for.
- traspler 5y agoRegarding macOS they write "On MacOS, as OpenGL is being deprecated, we will also supply builds based on ANGLE (over Metal this time) so the editor and exported games using GLES3 continues to run." but the ANGLE Readme says that GLES3.0 support is "in progress": https://chromium.googlesource.com/angle/angle/+/main/README.md https://chromium.googlesource.com/angle/angle/+/main/README.... with no status for all the following .X versions. Is the ANGLE Readme just outdated in regards to Metal support or am I missing something? This would also mean that Vulkan won't be an option when developing on macOS, right? Or will something like MoltenVK be shipped as well?
- sirn 5y agoVulkan on Godot on the Mac is using MoltenVK[1] [1]: https://twitter.com/reduzio/status/1215284328390889474 https://twitter.com/reduzio/status/1215284328390889474
- modeless 5y agoApple is implementing Metal support for ANGLE directly in the WebKit repo. The work is being upstreamed. The upstreaming process is not yet finished but Apple's implementation is fairly complete as it is the basis for the new WebGL 2 support in Safari. WebGL 2 is based on GLES3.0, so 3.1 or 3.2 features may not be supported (compute shaders being the most important feature there).