-
-
Notifications
You must be signed in to change notification settings - Fork 209
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Performance optimizations #11
Comments
# This is the 1st commit message: Parse gdextension_interface.h declarations using regex # This is the commit message #2: AsUninit trait to convert FFI pointers to their uninitialized versions # This is the commit message godot-rust#3: GodotFfi::from_sys_init() now uses uninitialized pointer types # This is the commit message godot-rust#4: Introduce GDExtensionUninitialized*Ptr, without changing semantics # This is the commit message godot-rust#5: Adjust init code to new get_proc_address mechanism # This is the commit message godot-rust#6: Make `trace` feature available in godot-ffi, fix interface access before initialization # This is the commit message godot-rust#7: Compatibility layer between Godot 4.0 and 4.1 (different GDExtension APIs) # This is the commit message godot-rust#8: Add GdextBuild to access build/runtime metadata # This is the commit message godot-rust#9: Detect 4.0 <-> 4.1 mismatches in both directions + missing `compatibility_minimum = 4.1` # This is the commit message godot-rust#10: Detect legacy/modern version of C header (also without `custom-godot` feature) # This is the commit message godot-rust#11: CI: add jobs that use patched 4.0.x versions # This is the commit message godot-rust#12: Remove several memory leaks by constructing into uninitialized pointers # This is the commit message godot-rust#13: CI: memcheck jobs for both 4.0.3 and nightly # This is the commit message godot-rust#14: Remove ToVariant, FromVariant, and VariantMetadata impls for pointers This commit splits SignatureTuple into two separate traits: PtrcallSignatureTuple and VarcallSignatureTuple. The latter is a child of the former. PtrcallSignatureTuple is used for ptrcall and only demands GodotFuncMarshall of its arguments. VarcallSignatureTuple is used for varcall and additionally demands ToVariant, FromVariant, and VariantMetadata of its arguments, so pointers cannot benefit from the optimizations provided by varcall over ptrcall. # This is the commit message godot-rust#15: Adds FromVariant and ToVariant proc macros # This is the commit message godot-rust#16: godot-core: builtin: reimplement Plane functions/methods # This is the commit message godot-rust#17: impl GodotFfi for Option<T> when T is pointer sized and nullable godot-rust#240 Additionally FromVariant and ToVariant are also implemented for Option<Gd<T>> to satisfy all the requirements for ffi and godot_api. # This is the commit message godot-rust#18: Fix UB in virtual method calls that take objects Fix incorrect incrementing of refcount when calling in to godot Fix refcount not being incremented when we receive a refcounted object in virtual methods # This is the commit message godot-rust#19: fix UB caused by preload weirdness # This is the commit message godot-rust#20: Implements swizzle and converts from/to tuples
caching fn generate_string_name() {
let foo = StringName::from("foo");
}
fn cached_string_name() {
static CACHED: OnceLock<StringName> = OnceLock::new();
let foo = CACHED.get_or_init(|| StringName::from("foo"));
} takes on average 0.15 seconds for the first one and 0.0003 seconds for the second one in release mode. we do make calls like this does require implementing |
Manually writing an implementation of
|
Caching the method bind, builtin method, and utility function pointers with a
|
The cached function pointers have been implemented, and with them, For anyone interested in deeper insights, I wrote about it here: |
I think we can close this, several things have been optimized and new ones probably deserve their own issue. |
Damn, impressive. I am ready to give these changes a try. Thanks everyone! |
There are some low-hanging fruits for making the library faster, this issue tracks such approaches.
Codegen
rustfmt
performance -- parallelize or allow skippingContext
, e.g.to_rust_type
Runtime
#[inline]
has a true impact (otherwise don't clutter the whole codebase)Object::call
) could useimpl AsVararg
+ passing as tuple, to allocate argument array on the stackThe text was updated successfully, but these errors were encountered: