We Keep Coding

OpenCV dll calls from Unity3D lead to FPS drop

I want to recognize ArUco Marker in Unity3D and attach a GameObject to their position. I know there are packages in the Asset Store, but as other people got it working I was looking for an free existing solution or try it by myself.
ArucoUnity The detection worked like a charm, but Unity crashes when accessing the data of the rvecs, tvecs. The error somewhere occur in the Get(int i) in the Vec3d class, when the c++ method au_cv_Vec3d_get(CppPtr, i, CppPtr); is called
OpenCv plus Unity This implementation seems not to be complete as there exist no function to EstimatePoseSingleMarker or similar to get the rvecs and tvecs. Also the last updated was Jan 2019.
As Unity (C#) provides the possibility to access unmanaged code, I followed this great tutorial and for the begging I was able to forward the cv::VideoCaputre stream to Unity. The only problem that occured was a FPS drop to around 35-40 whereas normally I get around 90-100.
The C# code:
void Update()
{
MatToTexture2D();
image.texture = tex;
}
void MatToTexture2D()
{
OpenCVInterop.GetRawImageBytes(pixelPtr);
//Update the Texture2D with array updated in C++
tex.SetPixels32(pixel32);
tex.Apply();
}
The C++ code:
extern "C" void __declspec(dllexport) __stdcall GetRawImageBytes(unsigned char* data)
{
_capture >> _currentFrame;
cv::Mat resizedMat(camHeight, camWidth, _currentFrame.type());
cv::resize(_currentFrame, resizedMat, resizedMat.size(), cv::INTER_CUBIC);
//Convert from RGB to ARGB
cv::Mat argb_img;
cv::cvtColor(resizedMat, argb_img, cv::COLOR_BGR2BGRA);
std::vector<cv::Mat> bgra;
cv::split(argb_img, bgra);
std::swap(bgra[0], bgra[3]);
std::swap(bgra[1], bgra[2]);
std::memcpy(data, argb_img.data, argb_img.total() * argb_img.elemSize());
}
The cause seems to be the first line _capture >> _currentFrame;, but as the others projects have to do the same (at least I guess so), I wonder if there is another reason.
If I don't manage to fix this issue I have to look for alternative approaches.

Just adding to / building on Mars' answer:
For the threading problem I would actually use a thread-save ConcurrentStack<Color32[]>. A stack is "last-in | first-out" so that the first item returned is always the last image data added by the thread.
The thread uses Push(pixel32) to add a new entry with image data
in Update you only use the latest entry (TryPop) for updating the texture.
The rest you ignore (Clear).
So something like
// the OpenCV thread will add(push) entries
// the Unity main thread will work on the entries
private ConcurrentStack<Color32[]> stack = new ConcurrentStack<Color32[]>();
public RawImage image;
public Texture2D tex;
private Thread thread;
void Start()
{
// Wherever you get your tex from
tex = new Texture2D(...);
// it should be enough to do this only once
// the texture stays the same, you only update its content
image.texture = tex;
}
// do things in OnEnable so everytime the object gets enabled start the thread
void OnEnable()
{
stack.Clear();
if(thread != null)
{
thread.Abort();
}
thread = new Thread(MatToTexture2D);
thread.Start();
}
void Update()
{
// here in the main thread work the stack
if (stack.TryPop(out var pixels32))
{
// Only use SetPixels and Apply when really needed
tex.SetPixels32(pixels32);
tex.Apply();
}
// Erase older data
stack.Clear();
}
// Make sure to terminate the thread everytime this object gets disabled
private void OnDisable()
{
if(thread == null) return;
thread.Abort();
thread = null;
}
// Runs in a thread!
void MatToTexture2D()
{
while(true)
{
try
{
// Do what you already have
OpenCVInterop.GetRawImageBytes(pixelPtr);
// However you convert the pixelPtr into Color32
Color32[] pixel32 = GetColorArrayFromPtr(pixelPtr);
// Now add this data to the stack
stack.Push(pixel32);
}
catch (ThreadAbortException ex)
{
// This exception is thrown when calling Abort on the thread
// -> ignore the exception since it is produced on purpose
}
}
}

If I recall correctly, the C++ call to get an image (_capture >> _currentFrame;) is blocking/synchronous, meaning your code won't continue until it actually retrieves the image. You probably want to run your MatToTexture2D code asynchronously.
※This will mean that your frame rate will be higher than your image retrieval rate.
Have your MatToTexture2D function run continuously as needed, updating tex. Then just continue to set your texture to the latest tex, which may be the same value 2-3 frames in a row.
Edit:
#derHugo's answer is much more solid for the programming side, so I'll hide that part. The basic issue is explained above, and derHugo's work-around is much better than my pseudo-code :)

Unity C# UI.Text issue: (...) This is not supported. UnityEngine.Canvas:SendWillRenderCanvases()

Trying to add LineDescriptor(Clone) (UnityEngine.UI.Text) for graphic rebuild while we are already inside a graphic rebuild loop. This is not supported. UnityEngine.Canvas:SendWillRenderCanvases()
Hey, I get that issue when I try to update positions of many game objects with Text Component attached. Have any idea what's the reason of that bug?
for(int i = 0; i< dottedLines.Count; i++)
{
dottedLineStaff[dottedLines[i]][1].transform.position = RectTransformUtility.WorldToScreenPoint(Camera.main, dottedLineStaff[dottedLines[i]][0].transform.position);
}

I encountered this error when I had an asynchronous plugin controlling a script that instantiates UI prefabs. Basically the plugin would occasionally instantiate a UI prefab during the Graphic.Rebuild loop. Parts of the UI prefab would draw incorrectly until I interacted with them.
My solution was to queue the instantiations for the next Update() call instead of instantiating them exactly when the plugin invoked a callback. Something similar to:
List<string> queue = new List<string>();
private object thisLock = new object();
public void PluginCallback(string name)
{
// instead of instantiating/modifying UI elements here, queue them for later
lock (thisLock)
{
queue.Add(name);
}
}
void Update()
{
lock (thisLock)
{
// shouldn't be in a Graphic.Rebuild loop now
if (queue.Count > 0)
{
foreach (string name in queue)
{
Text text = (Instantiate(prefab) as GameObject).GetComponent<Text>();
text.text = name;
}
queue.Clear();
}
}
}
Just FYI - I've also seen other posts where the issue was more likely a Unity bug (iOS, older Unity versions, etc.).
EDIT: added locking to example (https://msdn.microsoft.com/en-us/library/c5kehkcz.aspx). I'm pretty new to the concepts this issue involves, so I may be evolving my answer over time ... so, I give no claim to the example's general appropriateness overall - I'm only saying that in my case it resolved the errors the OP describes.

Waiting for event inside Unity Coroutine?

So I have a Unity coroutine method, in which I have some objects. These objects represent values that are being gathered from a server somewhere, and they send out an Updated event when they are ready.
I was wondering what the best way is to wait for all the values to be updated, inside a coroutine in Unity.
public IEnumerator DoStuff()
{
foreach(var val in _updateableValues)
{
if (val.Value != null) continue;
else **wait for val.Updated to be fired then continue**
}
//... here I do stuff with all the values
// I use the WWW class here, so that's why it's a coroutine
}
What would be the best way of doing something like this?
Thanks!

There is no builtin direct method to wait for the event itself, but you can use a synchronous nested coroutine to wait for a flag set by the event:
//Flag
bool eventHappened;
//Event subscriber that sets the flag
void OnEvent(){
eventHappened=true;
}
//Coroutine that waits until the flag is set
IEnumerator WaitForEvent() {
yield return new WaitUntil(eventHappened);
eventHappened=false;
}
//Main coroutine, that stops and waits somewhere within it's execution
IEnumerator MainCoroutine(){
//Other stuff...
yield return StartCoroutine(WaitForEvent());
//Oher stuff...
}
With that in mind, creating a generic coroutine that waits for an UnityEvent is easy:
private IEnumerator WaitUntilEvent(UnityEvent unityEvent) {
var trigger = false;
Action action = () => trigger = true;
unityEvent.AddListener(action.Invoke);
yield return new WaitUntil(()=>trigger);
unityEvent.RemoveListener(action.Invoke);
}

I thing that a better way is to check the sever every frame and not to wait for the an amount of time without any thinking.
public IEnumerator DoStuff()
{
/* wait until we have the value we want */
while( value != desiredValue)
yield return null
//after this loop, start the real processing
}
This is my little fix.

I use the WaitUntil directly where I need it. I found though, that WaitUntil doesn't want a boolean but expects a boolean predicate function - so I provide this in my example below.
The example is real running code from a game. It runs in the very first scene, where I start the music before I do - the slightly lengthly - loading of the other stuff.
//
public class AudioSceneInitializer : MonoBehaviour
{
[SerializeField] private GlobalEventsSO globalEvents;
//globalEvents is a scriptable object that - in my case - holds all events in the game
[SerializeField] private AudioManager audioManager;
//The audio manager which in my case will raise the audio ready event
[SerializeField] public GameObject audioGO;// contains AudioSources used by AudioManager
private bool audioReady = false;
// Start is called before the first frame update
IEnumerator Start()
{
if (audioManager != null)
//better check, if we don't produce the event, we wait forever
{
//subscribe to event
globalEvents.audioManagerReadyEvent += OnAudioReady;
//do something to raise the event - else we wait forever
audioManager.onInitialization(this);//"this" needed to hand over the audioGO to the AudioManager
// waits until the flag is set;
yield return new WaitUntil(IsAudioReady);
}
//now that we have music playing, we load the actual game
SceneManager.LoadSceneAsync(1, LoadSceneMode.Additive);
}
//Event subscriber that sets the flag
void OnAudioReady()
{
//unsubscribing, if - as in my case - it happens only once in the game
globalEvents.audioManagerReadyEvent -= OnAudioReady;
audioReady = true;
}
//predicate function that WaitUntil expects
private bool IsAudioReady()
{
return audioReady;
}
public void OnDisable()
{
audioManager.onCleanup();
}
}

A spin-lock is a solution, however not a very CPU-gentle one. In a spin-lock, you would just wait for the variable to have a certain value, otherwise sleep for a few milliseconds.
public IEnumerator DoStuff()
{
/* wait until we have the value we want */
while( value != desiredValue)
yield return new WaitForSeconds(0.001f);
//after this loop, start the real processing
}
Maybe you might want to think about restructuring your code, sothat no spin-lock is requiered, but a more interupt/event-based based approach can be implemented. That means, if you update a value and something has to take place after it happened, kick it off directly after changing that value. In C#, there's even an interface INotifyPropertyChanged for that design pattern (see MSDN), but you can easily design that yourself, too, e.g. by firing an event when that certain value has changed. We'd need more information on what exactly you want to react here, if you want a better solution than a spinlock, but this should give you some ideas.

How can i execute http request from not-main thread in unity? [duplicate]

My problem is I try to use Unity socket to implement something. Each time, when I get a new message I need to update it to the updattext (it is a Unity Text). However, When I do the following code, the void update does not calling every time.
The reason for I do not include updatetext.GetComponent<Text>().text = "From server: "+tempMesg;in the void getInformation is this function is in the thread, when I include that in getInformation() it will come with an error:
getcomponentfastpath can only be called from the main thread
I think the problem is I don't know how to run the main thread and the child thread in C# together? Or there maybe other problems.
Here is my code:
using UnityEngine;
using System.Collections;
using System;
using System.Net.Sockets;
using System.Text;
using System.Threading;
using UnityEngine.UI;
public class Client : MonoBehaviour {
System.Net.Sockets.TcpClient clientSocket = new System.Net.Sockets.TcpClient();
private Thread oThread;
// for UI update
public GameObject updatetext;
String tempMesg = "Waiting...";
// Use this for initialization
void Start () {
updatetext.GetComponent<Text>().text = "Waiting...";
clientSocket.Connect("10.132.198.29", 8888);
oThread = new Thread (new ThreadStart (getInformation));
oThread.Start ();
Debug.Log ("Running the client");
}
// Update is called once per frame
void Update () {
updatetext.GetComponent<Text>().text = "From server: "+tempMesg;
Debug.Log (tempMesg);
}
void getInformation(){
while (true) {
try {
NetworkStream networkStream = clientSocket.GetStream ();
byte[] bytesFrom = new byte[10025];
networkStream.Read (bytesFrom, 0, (int)bytesFrom.Length);
string dataFromClient = System.Text.Encoding.ASCII.GetString (bytesFrom);
dataFromClient = dataFromClient.Substring (0, dataFromClient.IndexOf ("$"));
Debug.Log (" >> Data from Server - " + dataFromClient);
tempMesg = dataFromClient;
string serverResponse = "Last Message from Server" + dataFromClient;
Byte[] sendBytes = Encoding.ASCII.GetBytes (serverResponse);
networkStream.Write (sendBytes, 0, sendBytes.Length);
networkStream.Flush ();
Debug.Log (" >> " + serverResponse);
} catch (Exception ex) {
Debug.Log ("Exception error:" + ex.ToString ());
oThread.Abort ();
oThread.Join ();
}
// Thread.Sleep (500);
}
}
}

Unity is not Thread safe, so they decided to make it impossible to call their API from another Thread by adding a mechanism to throw an exception when its API is used from another Thread.
This question has been asked so many times, but there have been no proper solution/answer to any of them. The answers are usually "use a plugin" or do something not thread-safe. Hopefully, this will be the last one.
The solution you will usually see on Stackoverflow or Unity's forum website is to simply use a boolean variable to let the main thread know that you need to execute code in the main Thread. This is not right as it is not thread-safe and does not give you control to provide which function to call. What if you have multiple Threads that need to notify the main thread?
Another solution you will see is to use a coroutine instead of a Thread. This does not work. Using coroutine for sockets will not change anything. You will still end up with your freezing problems. You must stick with your Thread code or use Async.
One of the proper ways to do this is to create a collection such as List. When you need something to be executed in the main Thread, call a function that stores the code to execute in an Action. Copy that List of Action to a local List of Action then execute the code from the local Action in that List then clear that List. This prevents other Threads from having to wait for it to finish executing.
You also need to add a volatile boolean to notify the Update function that there is code waiting in the List to be executed. When copying the List to a local List, that should be wrapped around the lock keyword to prevent another Thread from writing to it.
A script that performs what I mentioned above:
UnityThread Script:
#define ENABLE_UPDATE_FUNCTION_CALLBACK
#define ENABLE_LATEUPDATE_FUNCTION_CALLBACK
#define ENABLE_FIXEDUPDATE_FUNCTION_CALLBACK
using System;
using System.Collections;
using UnityEngine;
using System.Collections.Generic;
public class UnityThread : MonoBehaviour
{
//our (singleton) instance
private static UnityThread instance = null;
////////////////////////////////////////////////UPDATE IMPL////////////////////////////////////////////////////////
//Holds actions received from another Thread. Will be coped to actionCopiedQueueUpdateFunc then executed from there
private static List<System.Action> actionQueuesUpdateFunc = new List<Action>();
//holds Actions copied from actionQueuesUpdateFunc to be executed
List<System.Action> actionCopiedQueueUpdateFunc = new List<System.Action>();
// Used to know if whe have new Action function to execute. This prevents the use of the lock keyword every frame
private volatile static bool noActionQueueToExecuteUpdateFunc = true;
////////////////////////////////////////////////LATEUPDATE IMPL////////////////////////////////////////////////////////
//Holds actions received from another Thread. Will be coped to actionCopiedQueueLateUpdateFunc then executed from there
private static List<System.Action> actionQueuesLateUpdateFunc = new List<Action>();
//holds Actions copied from actionQueuesLateUpdateFunc to be executed
List<System.Action> actionCopiedQueueLateUpdateFunc = new List<System.Action>();
// Used to know if whe have new Action function to execute. This prevents the use of the lock keyword every frame
private volatile static bool noActionQueueToExecuteLateUpdateFunc = true;
////////////////////////////////////////////////FIXEDUPDATE IMPL////////////////////////////////////////////////////////
//Holds actions received from another Thread. Will be coped to actionCopiedQueueFixedUpdateFunc then executed from there
private static List<System.Action> actionQueuesFixedUpdateFunc = new List<Action>();
//holds Actions copied from actionQueuesFixedUpdateFunc to be executed
List<System.Action> actionCopiedQueueFixedUpdateFunc = new List<System.Action>();
// Used to know if whe have new Action function to execute. This prevents the use of the lock keyword every frame
private volatile static bool noActionQueueToExecuteFixedUpdateFunc = true;
//Used to initialize UnityThread. Call once before any function here
public static void initUnityThread(bool visible = false)
{
if (instance != null)
{
return;
}
if (Application.isPlaying)
{
// add an invisible game object to the scene
GameObject obj = new GameObject("MainThreadExecuter");
if (!visible)
{
obj.hideFlags = HideFlags.HideAndDontSave;
}
DontDestroyOnLoad(obj);
instance = obj.AddComponent<UnityThread>();
}
}
public void Awake()
{
DontDestroyOnLoad(gameObject);
}
//////////////////////////////////////////////COROUTINE IMPL//////////////////////////////////////////////////////
#if (ENABLE_UPDATE_FUNCTION_CALLBACK)
public static void executeCoroutine(IEnumerator action)
{
if (instance != null)
{
executeInUpdate(() => instance.StartCoroutine(action));
}
}
////////////////////////////////////////////UPDATE IMPL////////////////////////////////////////////////////
public static void executeInUpdate(System.Action action)
{
if (action == null)
{
throw new ArgumentNullException("action");
}
lock (actionQueuesUpdateFunc)
{
actionQueuesUpdateFunc.Add(action);
noActionQueueToExecuteUpdateFunc = false;
}
}
public void Update()
{
if (noActionQueueToExecuteUpdateFunc)
{
return;
}
//Clear the old actions from the actionCopiedQueueUpdateFunc queue
actionCopiedQueueUpdateFunc.Clear();
lock (actionQueuesUpdateFunc)
{
//Copy actionQueuesUpdateFunc to the actionCopiedQueueUpdateFunc variable
actionCopiedQueueUpdateFunc.AddRange(actionQueuesUpdateFunc);
//Now clear the actionQueuesUpdateFunc since we've done copying it
actionQueuesUpdateFunc.Clear();
noActionQueueToExecuteUpdateFunc = true;
}
// Loop and execute the functions from the actionCopiedQueueUpdateFunc
for (int i = 0; i < actionCopiedQueueUpdateFunc.Count; i++)
{
actionCopiedQueueUpdateFunc[i].Invoke();
}
}
#endif
////////////////////////////////////////////LATEUPDATE IMPL////////////////////////////////////////////////////
#if (ENABLE_LATEUPDATE_FUNCTION_CALLBACK)
public static void executeInLateUpdate(System.Action action)
{
if (action == null)
{
throw new ArgumentNullException("action");
}
lock (actionQueuesLateUpdateFunc)
{
actionQueuesLateUpdateFunc.Add(action);
noActionQueueToExecuteLateUpdateFunc = false;
}
}
public void LateUpdate()
{
if (noActionQueueToExecuteLateUpdateFunc)
{
return;
}
//Clear the old actions from the actionCopiedQueueLateUpdateFunc queue
actionCopiedQueueLateUpdateFunc.Clear();
lock (actionQueuesLateUpdateFunc)
{
//Copy actionQueuesLateUpdateFunc to the actionCopiedQueueLateUpdateFunc variable
actionCopiedQueueLateUpdateFunc.AddRange(actionQueuesLateUpdateFunc);
//Now clear the actionQueuesLateUpdateFunc since we've done copying it
actionQueuesLateUpdateFunc.Clear();
noActionQueueToExecuteLateUpdateFunc = true;
}
// Loop and execute the functions from the actionCopiedQueueLateUpdateFunc
for (int i = 0; i < actionCopiedQueueLateUpdateFunc.Count; i++)
{
actionCopiedQueueLateUpdateFunc[i].Invoke();
}
}
#endif
////////////////////////////////////////////FIXEDUPDATE IMPL//////////////////////////////////////////////////
#if (ENABLE_FIXEDUPDATE_FUNCTION_CALLBACK)
public static void executeInFixedUpdate(System.Action action)
{
if (action == null)
{
throw new ArgumentNullException("action");
}
lock (actionQueuesFixedUpdateFunc)
{
actionQueuesFixedUpdateFunc.Add(action);
noActionQueueToExecuteFixedUpdateFunc = false;
}
}
public void FixedUpdate()
{
if (noActionQueueToExecuteFixedUpdateFunc)
{
return;
}
//Clear the old actions from the actionCopiedQueueFixedUpdateFunc queue
actionCopiedQueueFixedUpdateFunc.Clear();
lock (actionQueuesFixedUpdateFunc)
{
//Copy actionQueuesFixedUpdateFunc to the actionCopiedQueueFixedUpdateFunc variable
actionCopiedQueueFixedUpdateFunc.AddRange(actionQueuesFixedUpdateFunc);
//Now clear the actionQueuesFixedUpdateFunc since we've done copying it
actionQueuesFixedUpdateFunc.Clear();
noActionQueueToExecuteFixedUpdateFunc = true;
}
// Loop and execute the functions from the actionCopiedQueueFixedUpdateFunc
for (int i = 0; i < actionCopiedQueueFixedUpdateFunc.Count; i++)
{
actionCopiedQueueFixedUpdateFunc[i].Invoke();
}
}
#endif
public void OnDisable()
{
if (instance == this)
{
instance = null;
}
}
}
USAGE:
This implementation allows you to call functions in the 3 most used Unity functions: Update, LateUpdate and FixedUpdate functions. This also allows you call run a coroutine function in the main Thread. It can be extended to be able to call functions in other Unity callback functions such as OnPreRender and OnPostRender.
1.First, initialize it from the Awake() function.
void Awake()
{
UnityThread.initUnityThread();
}
2.To execute a code in the main Thread from another Thread:
UnityThread.executeInUpdate(() =>
{
transform.Rotate(new Vector3(0f, 90f, 0f));
});
This will rotate the current Object the scipt is attached to, to 90 deg. You can now use Unity API(transform.Rotate) in another Thread.
3.To call a function in the main Thread from another Thread:
Action rot = Rotate;
UnityThread.executeInUpdate(rot);
void Rotate()
{
transform.Rotate(new Vector3(0f, 90f, 0f));
}
The #2 and #3 samples executes in the Update function.
4.To execute a code in the LateUpdate function from another Thread:
Example of this is a camera tracking code.
UnityThread.executeInLateUpdate(()=>
{
//Your code camera moving code
});
5.To execute a code in the FixedUpdate function from another Thread:
Example of this when doing physics stuff such as adding force to Rigidbody.
UnityThread.executeInFixedUpdate(()=>
{
//Your code physics code
});
6.To Start a coroutine function in the main Thread from another Thread:
UnityThread.executeCoroutine(myCoroutine());
IEnumerator myCoroutine()
{
Debug.Log("Hello");
yield return new WaitForSeconds(2f);
Debug.Log("Test");
}
Finally, if you don't need to execute anything in the LateUpdate and FixedUpdate functions, you should comment both lines of this code below:
//#define ENABLE_LATEUPDATE_FUNCTION_CALLBACK
//#define ENABLE_FIXEDUPDATE_FUNCTION_CALLBACK
This will increase performance.

I have been using this solution to this problem. Create a script with this code and attach it to a Game Object:
using System;
using System.Collections.Generic;
using System.Collections.Concurrent;
using UnityEngine;
public class ExecuteOnMainThread : MonoBehaviour {
public static readonly ConcurrentQueue<Action> RunOnMainThread = new ConcurrentQueue<Action>();
void Update()
{
if(!RunOnMainThread.IsEmpty)
{
while(RunOnMainThread.TryDequeue(out var action))
{
action?.Invoke();
}
}
}
}
Then when you need to call something on the main thread and access the Unity API from any other function in your application:
ExecuteOnMainThread.RunOnMainThread.Enqueue(() => {
// Code here will be called in the main thread...
});

Much of the writing about threads in Unity is incorrect.
How so?
Unity is, of course, totally frame-based.
When you work in a frame-based system, threading issues are completely different.
Threading issues on a frame-based system are completely different. (In fact, often much easier to deal with.)
Let's say you have a Unity thermometer display that shows some value
Thermo.cs
So it will have a function which is called in Update, like
func void ShowThermoValue(float fraction) {
display code to show the current thermometer value
}
Recall that the "Update" function in Unity simply means "run this once each frame".
That only runs once per frame, and that's that.
(Naturally, it runs only on the "main thread". There's nothing else in Unity! There's just ... "the Unity thread"!)
Somewhere else, perhaps in "IncomingData.cs", you will have a function which handles the concept "a new value has arrived":
[MonoPInvokeCallback(typeof(ipDel))]
public static void NewValueArrives(float f) {
... ???
}
Note that, of course, that is a class function! What else can it be?
You can't "reach in to" a normal Unity function. (Such as ShowThermoValue.) That would be meaningless - it's just a function which runs once each frame.Footnote 1
Let's say: values arrive very frequently and irregularly.
Image you have some sort of scientific devices (perhaps IR thermometers) connected connected to a rack of PCs
Those electronic devices deliver new "temperature" values very often. Let's say dozens of times per frame.
So, "NewValueArrives" is being called 100s of times a second.
So what do you do with the values?
It could not be simpler.
From the arriving-values thread, all you do is ................. wait for it ............. set a variable in the component!!
WTF? All you do is set a variable? That's it? How can it be that simple?
This is one of those unusual situations:
Much of the writing on threads in Unity is, simply, completely hopeless.
Surprisingly, the actual approach is extremely simple.
It's so simple you may think you are doing something wrong!!
So have the variable ...
[System.Nonserialized] public float latestValue;
Set it from the "arriving thread" ...
[MonoPInvokeCallback(typeof(ipDel))]
public static void NewValueArrives(float f) {
ThisScript.runningInstance.latestValue = f; // done
}
Honestly that's it.
Essentially, to be the world's greatest expert at "threading in Unity" - which is, obviously, frame-based - there's nothing more to do than the above.
And whenever ShowThermoValue is called each frame ...................... simply display that value!
Really, that's it!
[System.Nonserialized] public float latestValue;
func void ShowThermoValue() { // note NO arguments here!
display code, draws a thermometer
thermo height = latestValue
}
You're simply displaying the "latest" value.
latestValue may have been set once, twice, ten times, or a hundred times that frame ............ but, you simply display whatever is the value when ShowThermoValue runs that frame!
What else could you display?
The thermometer is updating at 60fps on-screen so you display the latest value. Footnote 2
It's actually that easy. It's just that easy. Surprising but true.
#(Critical aside - don't forget that vector3, etc, are NOT Atomic in Unity/C#)
As user #dymanoid has pointed out (read the important discussion below) it's critical to remember that while float is atomic in the Unity/C# milieu, anything else (say, Vector3 etc) IS NOT ATOMIC. Typically (as in the example here) you only pass floats around from calculations from, say, native plugins, thermometers etc. But it's essential to be aware that vectors and so on are NOT atomic.
Sometimes experienced threading programmers get in a knot with a frame-based system, because: in a frame based system most of the problems caused by racetrack and locking issues ... do not exist conceptually.
In a frame-based system, any game items should simply be displaying or behaving based on some "current value," which is set somewhere. If you have info arriving from other threads, just set those values - you're done.
You can not meaningfully "talk to the main thread" in Unity because that main thread ............. is frame-based!
Most locking, blocking and racetrack issues are non-existent in the frame-based paradigm because: if you set latestValue ten times, a million times, a billion times, in one particular frame .. what can you do? .. you can only display one value during that frame!
Think of an old-fashioned plastic film. You literally just have ...... a frame, and that's it. If you set latestValue a trillion times in one particular frame, ShowThermoValue will simply display (for that 60th of a second) the one value it grabs when it is run.
All you do is: leave information somewhere, which, the frame-paradigm system will utilize during that frame if it wants to.
That's it in a nutshell.
Thus, most "threading issues" disappear in Unity.
All you can do from
other calculation threads or
from plugin threads,
is just "drop-off values" which the game may use.
That's it!
Let's consider the question title...
How do you "... call a function in the main Thread"
This is completely meaningless. The "functions" in Unity are simply functions which the frame engine runs once per frame.
You can't "call" anything in Unity. The frame engine runs a number of things (many things) once per frame.
Note that indeed threads are totally irrelevant. If Unity ran with a billion threads, or with quantum computing, it would have no bearing on anything.
You can't "call a function" in a frame-based system.
Fortunately, the approach to take is dead simple, you just set values, which the frame-based functions can look at when they want! It's really that easy.
Footnotes
1 How could you? As a thought experiment, forget about the issue that you're on a different thread. ShowThermoValue is run once a frame by the frame engine. You can't "call" it in any meaningful way. Unlike in normal OO software, you cannot, say, instantiate an instance of the class (a Component?? meaningless) and run that function - that is completely meaningless.
In "normal" threaded programming, threads can talk back and fore and so on, and in doing so you have concerns about locking, racetrack and so on. But that is all meaningless in an ECS, frame-based, system. There is nothing to "talk to".
Let's say that Unity was in fact multithreaded!!!! So the Unity guys have all of the engine running in a multithreaded manner. It wouldn't make any difference - you can not get "in to" ShowThermoValue in any meaningful way! It's a Component which the frame engine runs once a frame and that's that.
So NewValueArrives is not anywhere - it's a class function!
Let's answer the question in the headline:
"Use Unity API from another Thread or call a function in the main Thread?"
The concept is >> completely meaningless <<. Unity (like all game engines) is frame-based. There's no concept of "calling" a function on the main thread. To make an analogy: it would be like a cinematographer in the celluloid-film era asking how to "move" something actually on one of the frames.
Of course that is meaningless. All you can do is change something for the next photo, the next frame.
2 I refer to the "arriving-values thread" ... in fact! NewValueArrives may, or may not, run on the main thread!!!! It may run on the thread of the plugin, or on some other thread! It may actually be completely single-threaded by the time you deal with the NewValueArrives call! It just doesn't matter! What you do, and all you can do, in a frame-based paradigm, is, "leave laying around" information which Components such as ShowThermoValue, may use, as, they see fit.

Another solution to run code on the main thread, but without requiring a game object and MonoBehavior, is to use SynchronizationContext:
// On main thread, during initialization:
var syncContext = System.Threading.SynchronizationContext.Current;
// On your worker thread
syncContext.Post(_ =>
{
// This code here will run on the main thread
Debug.Log("Hello from main thread!");
}, null);

Use UniRx's multithreading pattern, UniTask and RxSocket together.
[SerializeField] private Text m_Text;
async UniTaskVoid Connect() {
IPEndPoint endPoint = new IPEndPoint(IPAddress.IPv6Loopback, 12345);
// Create a socket client by connecting to the server at the IPEndPoint.
// See the UniRx Async tooling to use await
IRxSocketClient client = await endPoint.ConnectRxSocketClientAsync();
client.ReceiveObservable
.ToStrings()
.ObserveOnMainThread()
.Subscribe(onNext: message =>
{
m_Text.text = message;
}).AddTo(this);
// Send a message to the server.
client.Send("Hello!".ToByteArray());
}

How do I properly queue and execute tasks with variable start times in c#?

I'm still new to C#. Currently using .NET 3.5. I'm writing an online game, and my game engine needs to handle player requested unit movement. The units will move at different speeds and can have different destinations, so the actual time that the unit arrives (and the command completes) will differ widely.
My strategy is that the model will calculate the next move along a given path, and then queue a worker to execute around the same time the unit would be arriving at the location. The worker will update the unit, notify players, queue the next step, etc.
One of my biggest stumbling blocks was writing a worker thread that would conditionally execute a task. In the past my worker threads have just tried to run each task, in order, as quickly as possible. The other challenge is that unit moves might get queued that have earlier start times than any other task in the queue, so obviously, they would need to get moved to the front of the line.
This is what I have figured out so far, it works but it has some limitations, namely its 100ms sleep limits my command response time. Is there a pattern for what i'm trying to do, or am I missing something obvious?
Thank you!
public class ScheduledTaskWorker
{
List<ScheduledTask> Tasks = new List<ScheduledTask>();
public void AddTask(ScheduledTask task)
{
lock (Tasks)
{
Tasks.Add(task);
Tasks.Sort();
}
}
private void RemoveTask(ScheduledTask task)
{
lock (Tasks)
{
Tasks.Remove(task);
}
}
private ScheduledTask[] CopyTasks()
{
lock (Tasks)
{
return Tasks.ToArray();
}
}
public void DoWork()
{
while (!StopWorking)
{
ScheduledTask[] safeCopy = CopyTasks();
foreach (ScheduledTask task in safeCopy)
{
if (task.ExecutionTime > DateTime.Now)
break;
if (ThreadPool.QueueUserWorkItem(new WaitCallback(ThreadPoolCallBack), task))
RemoveTask(task);
else
{
// if work item couldn't be queued, stop queuing and sleep for a bit
break;
}
}
// sleep for 100 msec so we don't hog the CPU.
System.Threading.Thread.Sleep(100);
}
}
static void ThreadPoolCallBack(Object stateInfo)
{
ScheduledTask task = stateInfo as ScheduledTask;
task.Execute();
}
public void RequestStop()
{
StopWorking = true;
}
private volatile bool StopWorking;
}
And the task itself...
public class ScheduledTask : IComparable
{
Action<Item> Work;
public DateTime ExecutionTime;
Item TheItem;
public ScheduledTask(Item item, Action<Item> work, DateTime executionTime)
{
Work = work;
TheItem = item;
ExecutionTime = executionTime;
}
public void Execute()
{
Work(TheItem);
}
public int CompareTo(object obj)
{
ScheduledTask p2 = obj as ScheduledTask;
return ExecutionTime.CompareTo(p2.ExecutionTime);
}
public override bool Equals(object obj)
{
ScheduledTask p2 = obj as ScheduledTask;
return ExecutionTime.Equals(p2.ExecutionTime);
}
public override int GetHashCode()
{
return ExecutionTime.GetHashCode();
}
}

I'm not disagreeing with Chris, his answer is probably more appropriate to your underlying problem.
However, to answer your specific question, you could use something equivalent to Java's DelayQueue, which looks to have been ported to C# here: http://code.google.com/p/netconcurrent/source/browse/trunk/src/Spring/Spring.Threading/Threading/Collections/Generic/DelayQueue.cs?spec=svn16&r=16
Your Task class would implement IDelayed then you could either have a single thread calling Take (which blocks while there are no items ready) and executing task, or pass them off to ThreadPool.QueueUserWorkItem.

That sounds dangerously brittle. Without knowing more about your game and its circumstances, it sounds like you could potentially drown the computer in new threads just by moving a lot of units around and having them do long-running tasks. If this is server-side code and any number of players could be using this at once, that almost guarantees issues with too many threads. Additionally, there are timing problems. Complex code, bad estimating, delays in starting a given thread (which will happen with the ThreadPool) or a slowdown from a lot of threads executing simultaneously will all potentially delay the time when your task is actually executed.
As far as I know, the majority of games are single-threaded. This is to cut down on these issues, as well as get everything more in sync. If actions happen on separate threads, they may receive unfair prioritization and run unpredictably, thereby making the game unfair. Generally, a single game loop goes through all of the actions and performs them in a time-based fashion (i.e. Unit Bob can move 1 square per second and his sprite animates at 15 frames per second). For example, drawing images/animating, moving units, performing collision detection and running tasks would all potentially happen each time through the loop. This eliminates timing issues as well, because everything will get the same priority (regardless of what they're doing or how fast they move). This also means that the unit will know the instant that it arrives at its destination and can execute whatever task it's supposed to do, a little bit at a time each iteration.