Programming language benchmark? Measuring what? Programming seems so abstract and dependent on the programmer themselves that a benchmark seems strange to me. I'm not a programmer, though, so who knows?
Godot, huh. Isn't that a game engine? I hear it's been advancing along, but still has a fair way to go to compete with stuff like Unity and Unreal, but it has that indie-ish charm to it that those don't.

>>119241
Are you seriously waiting for Godot?

>>119221
Godot has a repo for benchmarks: https://github.com/godotengine/godot-benchmarks
There's a few in there specific to GDScript, such as memory allocation or string manipulation. I'd recommend looking into contributing your own benchmarks if you want to flesh it out, since stuff like dictionary manipulation or simple number crunching seems absent.

>How do I make good programming language benchmarks
The only real requirement for a benchmark is to be consistent between runs and to be able to highlight small differences in performance. You could design one for isolating a known weak point in a language's runtime, or one that can be compared to what someone in an actual use case might do. It just depends on what you're trying to measure.

>Any benchmarks only look at framerate in physics calcd and not GC impact or memory per object
The problem is there is no GC in GDScript like you would find in Python or JS. Objects are instead reference counted, so deallocation occurs as soon as no other object references it, rather than waiting for the GC to do a cycle. In essence, this means there are no random GC pauses to deal with, but you run the risk of creating a reference cycle that can't be automatically deallocated, or having objects deallocated at a bad time and causing a frame stutter.

I also suspect most game developers would rather spend time profiling their specific project than profiling the language itself, since there's always the option to drop down to C++/C# if needed.