Install Asset
Install via Godot
To maintain one source of truth, Godot Asset Library is just a mirror of the old asset library so you can download directly on Godot via the integrated asset library browser
Quick Information
Dispatch Queue implementation.Supports synchronous and threaded execution, with the later being either serial (1 Thread) or concurrent (2+ Threads, a.k.a. Thread Pool).If threading is not supported by OS, fallbacks to running in synchronous mode.There are signals for when each task is finished and another for when all tasks are finished, so it's very easy to hook callbacks or yield until they are emitted.You can also dispatch a group of tasks at once and respond to a signal when all of them are finished.Dispatch queues are References, but there are Node and Resource wrappers for it to ease integrating with other nodes and scenes.
Dispatch Queue
Threaded or synchronous Dispatch Queues for Godot.
Threaded Dispatch Queues are also known as Thread Pools.
Available at the Asset Library.
Usage
const DispatchQueue = preload("res://addons/dispatch_queue/dispatch_queue.gd")
# 1) Instantiate
var dispatch_queue = DispatchQueue.new()
# 2.a) Either create a serial...
dispatch_queue.create_serial()
# 2.b) ...or concurrent queue
dispatch_queue.create_concurrent(OS.get_processor_count())
# (if you do neither, DispatchQueue will run in synchronous mode)
# 3) Dispatch methods, optionally responding to tasks and task groups "finished" signal
# 3.a) Fire and forget style
dispatch_queue.dispatch(self, "method_name", ["optional", "method", "arguments"]).then(self, "result_callback")
dispatch_queue.dispatch_group([
[self, "method_name1", ["optional", "arguments"]],
[self, "method_name2"],
[self, "method_name3"],
]).then_deferred(self, "group_results_callback")
# 3.b) Coroutine style
var task = dispatch_queue.dispatch(self, "mymethod")
var mymethod_result = yield(task, "finished")
var task_group = dispatch_queue.dispatch_group([ [self, "method1"], [self, "method2"] ])
var group_method_results = yield(task_group, "finished")
# 4) Optionally respond to the `all_tasks_finished` signal to know when all tasks have finished
# 4.a) Connect style
dispatch_queue.connect("all_tasks_finished", self, "_on_all_tasks_finished")
# 4.b) Coroutine style
yield(dispatch_queue, "all_tasks_finished")
# DispatchQueue extends Reference, so no need to worry about freeing it manually
There is a Node script (addons/dispatch_queue/dispatch_queue_node.gd) that wraps every aspect of dispatch queues. Useful for having a local queue in a scene or as an Autoload.
There is also a Resource script (addons/dispatch_queue/dispatch_queue_resource.gd) that wraps every aspect of dispatch queues. Useful for sharing queues with multiple objects between scenes without resorting to Autoload.
API
DispatchQueue (addons/dispatch_queue/dispatch_queue.gd):
signal all_tasks_finished()
- Emitted when the last queued Task finishes. This signal is emitted deferred, so it is safe to call non Thread-safe APIs.
create_serial()
- Creates a Thread of execution to process tasks.
If threading is not supported, fallback to synchronous mode.
If queue was already serial, this is a no-op, otherwise
calls
shutdown
and create a new Thread.
create_concurrent(thread_count: int = 1)
- Creates
thread_count
Threads of execution to process tasks. If threading is not supported, fallback to synchronous mode. If queue was already concurrent withthread_count
Threads, this is a no-op, otherwise callsshutdown
and create new Threads. Ifthread_count <= 1
, creates a serial queue.
dispatch(object: Object, method: String, args: Array = []) -> Task
- Create a Task for calling
method
onobject
withargs
. On threaded mode, the Task will be executed on a Thread when there is one available. On synchronous mode, the Task will be executed on the next frame.
dispatch_group(task_list: Array) -> TaskGroup
- Create all tasks in
task_list
by callingdispatch
on each value, returning the TaskGroup associated with them.task_list
should be an Array of Arrays, each of them containing the object, method and optional args Array, in this order.
is_threaded() -> bool
- Returns whether queue is threaded or synchronous.
get_thread_count() -> int
- Returns the current Thread count. Returns 0 on synchronous mode.
size() -> int
- Returns the number of queued tasks.
is_empty() -> bool
- Returns whether queue is empty, that is, there are no tasks queued.
clear()
- Cancel pending Tasks, clearing the current queue.
shutdown()
- Cancel pending Tasks, wait and release the used Threads.
This method is called automatically on
NOTIFICATION_PREDELETE
. It is safe to call this more than once.
Task (inner class of DispatchQueue)
signal finished(result)
- Emitted after Task executes its method, passing the result as argument.
The signal is emitted in the same Thread that executed the Task, so you
need to connect with
CONNECT_DEFERRED
if you want to call non Thread-safe APIs.
then(signal_responder: Object, method: String, binds: Array = [], flags: int = 0)
- Helper method for connecting to the "finished" signal.
Always adds
CONNECT_ONESHOT
to flags. This enables the following pattern:
dispatch_queue.dispatch(object, method).then(signal_responder, method)
then_deferred(signal_responder: Object, method: String, binds: Array = [], flags: int = 0)
- Alias for
then
that also addsCONNECT_DEFERRED
to flags.
TaskGroup (inner class of DispatchQueue)
signal finished(results)
- Emitted after all Tasks in the group finish, passing the results Array as argument.
The signal is emitted in the same Thread that executed the last pending Task, so you
need to connect with
CONNECT_DEFERRED
if you want to call non Thread-safe APIs.
then(signal_responder: Object, method: String, binds: Array = [], flags: int = 0)
- Helper method for connecting to the "finished" signal.
Always adds
CONNECT_ONESHOT
to flags. This enables the following pattern:
dispatch_queue.dispatch_group(task_list).then(signal_responder, method)
then_deferred(signal_responder: Object, method: String, binds: Array = [], flags: int = 0)
- Alias for
then
that also addsCONNECT_DEFERRED
to flags.
DispatchQueueNode (addons/dispatch_queue/dispatch_queue_node.gd):
Node that wraps a DispatchQueue.
Apart from creation, all DispatchQueue public methods and signals are supported.
Creates the Threads when entering tree and shuts down when exiting tree.
export(int) var thread_count = -1
- Number of Threads DispatchQueue will utilize.
If
thread_count == 0
, runs queue in synchronous mode. Ifthread_count < 0
, createsOS.get_processor_count()
Threads.
DispatchQueueResource (addons/dispatch_queue/dispatch_queue_resource.gd):
Resource that wraps a DispatchQueue.
Apart from creation, all DispatchQueue public methods and signals are supported.
export(int) var thread_count = -1
- Number of Threads DispatchQueue will utilize.
If
thread_count == 0
, runs queue in synchronous mode. Ifthread_count < 0
, createsOS.get_processor_count()
Threads.
Credits
- Conveyor icon by smalllikeart: https://www.flaticon.com/free-icon/conveyor_888545
Similar projects
- GODOThreadPOOL: https://github.com/zmarcos/godothreadpool
Dispatch Queue implementation.
Supports synchronous and threaded execution, with the later being either serial (1 Thread) or concurrent (2+ Threads, a.k.a. Thread Pool).
If threading is not supported by OS, fallbacks to running in synchronous mode.
There are signals for when each task is finished and another for when all tasks are finished, so it's very easy to hook callbacks or yield until they are emitted.
You can also dispatch a group of tasks at once and respond to a signal when all of them are finished.
Dispatch queues are References, but there are Node and Resource wrappers for it to ease integrating with other nodes and scenes.
Reviews
Quick Information
Dispatch Queue implementation.Supports synchronous and threaded execution, with the later being either serial (1 Thread) or concurrent (2+ Threads, a.k.a. Thread Pool).If threading is not supported by OS, fallbacks to running in synchronous mode.There are signals for when each task is finished and another for when all tasks are finished, so it's very easy to hook callbacks or yield until they are emitted.You can also dispatch a group of tasks at once and respond to a signal when all of them are finished.Dispatch queues are References, but there are Node and Resource wrappers for it to ease integrating with other nodes and scenes.