Critical Sections FAQ

1. When is it not mandatory to protect by a mutex one shared variable between several threads?
2. What is the chronology of code execution of 2 critical sections (with a mutex locking and a conditional variable signaling) that compete between 2 threads?
3. What happens if calling 'Condsignal()' or 'Condbroadcast()' without mutex locked?
4. Why is it mandatory to put 'Condwait' within a 'While...Wend' loop for checking a Boolean predicate (set by other thread before activate 'Condsignal' or 'Condbroadcast', and reset after the 'Wend')?

4.1. Is 'Condwait' (and 'Condsignal' or 'Condbroadcast') still useful when there is already a 'While' loop for checking a Boolean predicate set by other thread?

5. How to implement a user input-line function fully thread-safe?
6. How to use 'Screenlock' with multi-threading?
7. How to use 'video paging (double buffering or page flipping)' with multi-threading?
8. How to use the FB runtime library for multi-threaded applications (gfxlib2) with multi-threading?
9. How to use console statements and keyboard inputs with multi-threading?
10. Is it better to take precautions when using the keyword 'Sleep' in threads?
11. Can all tools to handle multi-threading be encapsulated in a base Class (that the user extends with a derived Type for his own implementing)?
12. What is the execution delay of the code of a thread after the thread is created by 'ThreadCreate'?
13. What is the synchronization latency when synchronizing threads either by mutual exclusions or by conditional variables?
14. What happens when multiple threads are waiting on the same condition variable?
15. How to optimize sequencing of successive user tasks executed by threading?
16. Why is multi-threading performance penalized by many shared memory accesses (even more in writing mode)?
17. Why is multi-threading performance heavily penalized by many manipulation of var-len strings and var-len arrays?

When accessing to shared variables between several threads, all their accesses must be generally put inside blocks Mutexlock...Mutexunlock, in all threads:

- When the shared variable is only one simple predefined numeric type of size <= sizeof(integer) (only one assembler instruction for access), the mutex use may be not mandatory.
- But if this is for example one shared variable LongInt with a win32 compilation, it is advised here to use a mutex (otherwise the reading phase by a thread may be interlaced with the writing phase of another thread).

That is because to access a variable in memory (for reading or for writing), a processor uses its internal registers.
A N-bit processor has N-bit registers but none greater:

- So one only assembler instruction allows it to access a N-bit variable in memory.
- At opposite, to access a 2N-bit variable, it must use 2 assembler instructions.
- If between these two assembler instructions (for writing), another thread accesses this same variable (for reading), the got value may be incoherent (N-bit highest and N-bit lowest incoherent together).

This behavior can be checked with a graphic program using two threads and a shared LongInt (64-bit) without mutex:

- by compiling in 32-bit, many read values are incoherent.
- by compiling in 64-bit, no read value is incoherent.

Compile the below test program:

- in 32-bit => Many erroneous points not on the circle but anywhere in the square containing the circle. If you uncomment the four lines 37/39/58/60 to activate the mutex, then all the got points are now on the circle only.
- in 64-bit => All points are valid, on the circle only, even if the mutex is not activated.

' - The "user-defined thread" computes the points coordinates on a circle,
' and write those in a LongInt (32-bit & 32-bit = 64-bit)
' - The "main thread" plots the points from the LongInt value.
'
' Behavior:
' - The first point must be pre-determined.
' - Nothing prevents that a same calculated point could be plotted several times
' (depends on execution times of the loops between main thread and user thread).
' - Nothing prevents that a calculated point could be not plotted
' (same remark on the loop times).
'
' Remark:
' Voluntarily, there is no Sleep in the loop of each thread (normally strongly discouraged),
' but this is just in this special case to amplify the behavior effects to observe.

Union Point2D
Dim As LongInt xy
Type
Dim As Long y
Dim As Long x
End Type
End Union

Dim As Any Ptr handle
Dim Shared As Any Ptr mutex
Dim Shared As Integer quit

Sub Thread (ByVal param As Any Ptr)
Const pi As Single = 4 * Atn(1)
Dim As Point2D Ptr p = param
Do
Dim As Point2D P2D0
Dim As Single teta = 2 * pi * Rnd
P2D0.x = 320 + 200 * Cos(teta)
P2D0.y = 240 + 200 * Sin(teta)
' Mutexlock(mutex)
p->xy = P2D0.xy
' Mutexunlock(mutex)
' Sleep 5, 1
Loop Until quit = 1
End Sub

Screen 12

Dim As Point2D P2D
P2D.x = 520
P2D.y = 240

mutex = MutexCreate
handle = ThreadCreate(@Thread, @P2D)

Dim As Integer c

Do
Dim As Point2D P2D0
' Mutexlock(mutex)
P2D0.xy = P2D.xy
' Mutexunlock(mutex)
PSet (P2D0.x, P2D0.y), c
c = (c Mod 15) + 1
' Sleep 5, 1
Loop Until Inkey <> ""

quit = 1
ThreadWait(handle)
MutexDestroy(mutex)

Chronology for one thread signaling which occurs:

a) while another thread is waiting (within a While loop on predicate),
b) before another thread is waiting (within a While loop on predicate).

#define while_loop_on_predicate

Dim As Any Ptr handle
Dim Shared As Any Ptr mutex
Dim Shared As Any Ptr cond
Dim As Integer sleep0
Dim As Integer sleep1
#ifdef while_loop_on_predicate
Dim Shared As Integer ready
#endif

Sub Thread1 (ByVal param As Any Ptr)
Sleep *Cast(Integer Ptr, param), 1
MutexLock(mutex)
Color 11 : Print " Thread#1 locks the mutex"
Color 11 : Print " Thread#1 executes code with exclusion"
#ifdef while_loop_on_predicate
ready = 1
#endif
Color 11 : Print " Thread#1 is signaling"
CondSignal(cond)
Color 11 : Print " Thread#1 executes post-code with exclusion"
Color 11 : Print " Thread#1 unlocks the mutex"
MutexUnlock(mutex)
End Sub

Sub Thread0 (ByVal param As Any Ptr)
Sleep *Cast(Integer Ptr, param), 1
MutexLock(mutex)
Color 10 : Print " Thread#0 locks the mutex"
Color 10 : Print " Thread#0 executes pre-code with exclusion"
#ifdef while_loop_on_predicate
While ready <> 1
#endif
Color 10 : Print " Thread#0 is waiting"
CondWait(cond, mutex)
Color 10 : Print " Thread#0 is waked"
#ifdef while_loop_on_predicate
Wend
#endif
Color 10 : Print " Thread#0 executes code with exclusion"
#ifdef while_loop_on_predicate
ready = 0
#endif
Color 10 : Print " Thread#0 unlocks the mutex"
MutexUnlock(mutex)
End Sub

mutex = MutexCreate
cond = CondCreate

sleep0 = 0
sleep1 = 1000
Color 7 : Print "Chronology for Thread#1 signaling while Thread#0 is waiting:"
handle = ThreadCreate(@Thread1, @sleep1)
Thread0(@sleep0)
ThreadWait(handle)
Color 7 : Print "Thread#1 finished": Print
Sleep 1000, 1

sleep0 = 1000
sleep1 = 0
Color 7 : Print "Chronology for Thread#1 signaling before Thread#0 is waiting:"
handle = ThreadCreate(@Thread1, @sleep1)
Thread0(@sleep0)
ThreadWait(handle)
Color 7 : Print "Thread#1 finished": Print

MutexDestroy(mutex)
CondDestroy(cond)
Sleep

Output part a - Chronology for Thread#1 signaling while Thread#0 is waiting:

Chronology for Thread#1 signaling while Thread#0 is waiting:
	Thread#0 locks the mutex
	Thread#0 executes pre-code with exclusion
	Thread#0 is waiting
		Thread#1 locks the mutex
		Thread#1 executes code with exclusion
		Thread#1 is signaling
		Thread#1 executes post-code with exclusion
		Thread#1 unlocks the mutex
	Thread#0 is waked
	Thread#0 executes code with exclusion
	Thread#0 unlocks the mutex
Thread#1 finished

Output part b - Chronology for Thread#1 signaling before Thread#0 is waiting:

Chronology for Thread#1 signaling before Thread#0 is waiting:
		Thread#1 locks the mutex
		Thread#1 executes code with exclusion
		Thread#1 is signaling
		Thread#1 executes post-code with exclusion
		Thread#1 unlocks the mutex
	Thread#0 locks the mutex
	Thread#0 executes pre-code with exclusion
	Thread#0 executes code with exclusion
	Thread#0 unlocks the mutex
Thread#1 finished

Note: If CondWait is not within a While loop on predicate (by putting in comment the first line of above program), one can check in the second case (thread#1 signaling before thread#0 waiting), that thread#0 remains blocked in its waiting phase (Ctrl-C to quit).

Referring to the example 2 on the Critical Sections, one takes this opportunity to recall that:

- The mutex must always be also locked while executing Condsignal() or Condbroadcast() to wake up a thread (it may be unlocked but only after Condsignal() or Condbroadcast()).
- If the mutex is not locked (or even if the mutex is unlocked only just before executing Condsignal() or Condbroadcast()), the behavior may become unpredictable (it may work or not, depending on the threads configuration and execution real time).

In the example 2 on the Critical Sections "Synchronous method example using a condwait then a condbroadcast (and a mutex) for all threads":

- If one at least Mutexunlock() is moved just before its Condbroadcast(), the program hangs very quickly.
- Although some users certify that the mutex can always be unlocked just before Condsignal() or Condbroadcast(), and others more cautious assert that one can do it only for a Condbroadcast(), experiment shows the opposite!

The general rule is that:

- The condition must not be signaled (by Condsignal() or Condbroadcast()) between the time a thread locks the mutex and the time it waits on the condition variable (CondWait()), otherwise it seems that it may damage the waiting queue of threads on that condition variable.
- Thus to avoid that and follow this rule, it is necessary that the mutex remains locked when the condition is signaled.

While predicate <> True Condwait(conditionalid, mutexid) Wend predicate = False

In all documentations, it is highly advisable to do so, mainly justified to fight against eventual spurious wake-ups.

This is probably true, but it is also advisable to do so to avoid to loose a CondSignal (or CondBroadcast) if it is prematurely activated while the receiving thread is not yet waiting on CondWait (the signal is lost forever):

- In that case, the receiving thread has even not yet locked the mutex before that CondSignal (or CondBroadcast) is activated.
- So the predicate will already true before the receiving thread reaches the 'While...Wend' loop, inducing that CondWait is downright skipped, so avoiding a definitive blocking phenomenon.

Let two threads (thread #0 in main program, thread #1 in a user procedure, each that prints its number in a loop), having about the same execution time, and each one synchronizing the other in order to well interlace their numbers (by using one mutex, two condition variables and CondSignal/CondWait):

Without a 'While...Wend' loop on predicate, the program hangs quickly (Ctrl-C to quit):

'            Thread#0                           XOR + <==>                       Thread#1
'   .....                                                             .....
'   MutexLock(mut) <-------------------------.             .----------> MutexLock(mut)
'   ( atomic_mutex_unlock(mut) ) ------.     |             |            Do_something_with_exclusion
'   CondWait(cond#1, mut) <----------- | --- | ----------- | ---------- CondSignal(cond#1)
'   ( atomic_mutex_re-lock(mut) ) <--- | ----'----.        |     .----- ( atomic_mutex_unlock(mut) )
'   Do_something_with_exclusion        |     .--- | ------ | --- | ---> CondWait(cond#2, mut)
'   CondSignal(cond#2) --------------- | ----'    |    .---'---- | ---> ( atomic_mutex_re-lock(mut) )
'   Do_something_with_exclusion        |     .--- | ---'         |      Do_something_with_exclusion
'   MutexUnlock(mut) ------------------'-----'    '--------------'----- MutexUnlock(mut)
'   .....                                                               .....
'
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)

Dim As Any Ptr handle
Dim Shared As Any Ptr mutex
Dim Shared As Any Ptr cond1
Dim Shared As Any Ptr cond2
Dim Shared As Integer quit

Sub Thread (ByVal param As Any Ptr)
Do
MutexLock(mutex)
Print "1";
CondSignal(cond1)
CondWait(cond2, mutex)
If quit = 1 Then
MutexUnlock(mutex)
Exit Do
End If
MutexUnlock(mutex)
Sleep 1, 1
Loop
End Sub

mutex = MutexCreate
cond1 = CondCreate
cond2 = CondCreate
handle = ThreadCreate(@Thread)

Do
MutexLock(mutex)
CondWait(cond1, mutex)
Print "0";
CondSignal(cond2)
If Inkey <> "" Then
quit = 1
MutexUnlock(mutex)
Exit Do
End If
MutexUnlock(mutex)
Sleep 1, 1
Loop

ThreadWait(handle)
MutexDestroy(mutex)
CondDestroy(cond1)
CondDestroy(cond2)
Print

Sleep

With a 'While...Wend' loop on predicate around each CondWait, no blocking phenomenon:

'            Thread#0                                 XOR + <==>                            Thread#1
'   .....                                                                          .....
'   MutexLock(mut) <-----------------------------.                          .----> MutexLock(mut)
'   While bool#1 <> true <---------------------- | --------.                |      Do_something_with_exclusion
'       ( atomic_mutex_unlock(mut) ) ------.     |         '--------------- | ---- bool#1 = true
'       CondWait(cond#1, mut) <----------- | --- | ------------------------ | ---- CondSignal(cond#1)
'       ( atomic_mutex_re-lock(mut) ) <--- | ----'----.    .--------------- | ---> While bool#2 <> true
'   Wend                                   |          |    |          .---- | -------- ( atomic_mutex_unlock(mut) )
'   bool#1 = false                .------- | -------- | ---'  .------ | --- |--------> CondWait(cond#2, mut)
'   Do_something_with_exclusion   |   .--- | -------- | ------'  .--- | ----'--------> ( atomic_mutex_re-lock(mut) )
'   bool#2 = true ----------------'   |    |     .--- | ---------'    |            Wend
'   CondSignal(cond#2) ---------------'    |     |    |               |            bool#2 = false
'   Do_something_with_exclusion            |     |    |               |            Do_something_with_exclusion
'   MutexUnlock(mut) ----------------------'-----'    '---------------'----------- MutexUnlock(mut)
'   .....                                                                          .....
'
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)

Dim As Any Ptr handle
Dim Shared As Any Ptr mutex
Dim Shared As Any Ptr cond1
Dim Shared As Any Ptr cond2
Dim Shared As Integer new1
Dim Shared As Integer new2
Dim Shared As Integer quit

Sub Thread (ByVal param As Any Ptr)
Do
MutexLock(mutex)
Print "1";
new1 = 1
CondSignal(cond1)
While new2 <> 1
CondWait(cond2, mutex)
Wend
new2 = 0
If quit = 1 Then
MutexUnlock(mutex)
Exit Do
End If
MutexUnlock(mutex)
Sleep 1, 1
Loop
End Sub

mutex = MutexCreate
cond1 = CondCreate
cond2 = CondCreate
handle = ThreadCreate(@Thread)

Do
MutexLock(mutex)
While new1 <> 1
CondWait(cond1, mutex)
Wend
new1 = 0
Print "0";
new2 = 1
CondSignal(cond2)
If Inkey <> "" Then
quit = 1
MutexUnlock(mutex)
Exit Do
End If
MutexUnlock(mutex)
Sleep 1, 1
Loop

ThreadWait(handle)
MutexDestroy(mutex)
CondDestroy(cond1)
CondDestroy(cond2)
Print

Sleep

4.1. Is 'Condwait' (and 'Condsignal' or 'Condbroadcast') still useful when there is already a 'While...Wend' loop for checking a Boolean predicate set by other thread?

(another question induced by the previous one)

- The recommended structure is as follows:

'  Principle of mutual exclusion + CONDWAIT in a While...Wend loop with predicate check, for a thread sub-section
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)
'
'
'          Thread                                        Other Thread
'      MUTEXLOCK(mutexID) <------------------------- from ( atomic_mutex_unlock(mutexID) ) or MUTEXUNLOCK(mutexID)
'      .......
'      While booleanT <> True <--------------------- from booleanT = True
'          ( atomic_mutex_unlock(mutexID) ) -------> to MUTEXLOCK(mutexID) or ( atomic_mutex_re-lock(mutexID) )
'          CONDWAIT(conditionalID, mutexID) <------- from CONDSIGNAL(conditionalID)
'          ( atomic_mutex_re-lock(mutexID) ) <------ from ( atomic_mutex_unlock(mutexID) ) or MUTEXUNLOCK(mutexID)
'      Wend
'      booleanT = False
'      .......
'      MUTEXUNLOCK(mutexID) -----------------------> to MUTEXLOCK(mutexID) or ( atomic_mutex_re-lock(mutexID) )

'  Principle of mutual exclusion + CONDSIGNAL with predicate check, for a thread sub-section
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)
'
'          Thread                                        Other Thread
'      MUTEXLOCK(mutexID) <------------------------- from ( atomic_mutex_unlock(mutexID) ) or MUTEXUNLOCK(mutexID)
'      .......
'      booleanOT = True ---------------------------> to While booleanOT <> True
'      CONDSIGNAL(conditionalID) ------------------> to CONDWAIT(conditionalID, mutexID)
'      .......
'      MUTEXUNLOCK(mutexID) -----------------------> to MUTEXLOCK(mutexID) or ( atomic_mutex_re-lock(mutexID) )

- If 'CondWait' is not used, it is mandatory to put instead a 'Sleep x, 1' instruction in the 'While...Wend' loop on the Boolean flag, in order to release time-slice when looping (in addition, this 'Sleep x, 1' must be put inside a ['Mutexunlock'...'Mutexlock'] block to release another thread):

'  Principle of mutual exclusion + SLEEP in a While...Wend loop with predicate check, for a thread sub-section
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)
'
'          Thread                                        Other Thread
'      MUTEXLOCK(mutexID) <------------------------- from MUTEXUNLOCK(mutexID)
'      .......
'      While booleanT <> True <--------------------- from booleanT = True
'          MUTEXUNLOCK(mutexID) -------------------> to MUTEXLOCK(mutexID)
'          SLEEP(tempo, 1)
'          MUTEXLOCK(mutexID) <--------------------- from MUTEXUNLOCK(mutexID)
'      Wend
'      booleanT = False
'      .......
'      MUTEXUNLOCK(mutexID) -----------------------> to MUTEXLOCK(mutexID)

'  Principle of mutual exclusion + predicate check only, for a thread sub-section
'  (connecting lines join the sender(s) and receiver(s) impacted by each action occurring during the sequence)
'
'          Thread                                        Other Thread
'      MUTEXLOCK(mutexID) <------------------------- from MUTEXUNLOCK(mutexID)
'      .......
'      booleanOT = True ---------------------------> to While booleanOT <> True
'      .......
'      MUTEXUNLOCK(mutexID) -----------------------> to MUTEXLOCK(mutexID)

During 'CondWait', the thread execution is suspended and does not consume any CPU time until the condition variable is signaled.
But if 'Sleep x, 1' is put instead, the waiting time is predetermined and not self adaptive like that of 'CondWait'.

=> 'CondWait' is useful to optimize the execution time.

The Input keyword may be not thread-safe, when another thread must also access to input/output resource:

- When executing the Input statement, the other running threads must not change the position of the text cursor, which prohibits instructions such as Locate, Print, ...
- Moreover, one cannot enclosed the Input keyword inside a mutex locking (as we can do it for the Inkey keyword), because while the inputting line would be not completed and validated, the other threads that want to also access to input/output would be fully blocked (waiting for mutex unlocking).

Thread-safe input-line function (versus input/output resource):

Input position, prompt message, sleeping time, line-blanking command, mutex pointer can be passed to the following threadInput() function that simulates a simplified input function, but thread-safe, by using a looping around the Inkey keyword (all input/output keywords must be enclosed inside a mutex locking block, and the cursor position must be restored at each mutex locking block ending):

Function threadInput (ByVal row As Integer, ByVal column As Integer, ByRef prompt As String = "", _
ByVal sleeptime As Integer = 15, ByVal blank As Integer = 0, ByVal mutex As Any Ptr = 0 _
) As String
Dim As String inputchr
Dim As String inputline
Dim As Integer cursor
Dim As Integer cursor0
Dim As Integer r
Dim As Integer c

MutexLock(mutex)
r = CsrLin()
c = Pos()
Locate row, column
Print prompt & " _";
cursor0 = Pos() - 1
Locate r, c
MutexUnlock(mutex)

Do
MutexLock(mutex)
r = CsrLin()
c = Pos()
inputchr = Inkey
If inputchr <> "" Then
If inputchr >= Chr(32) And inputchr < Chr(255) Then
inputline = Left(inputline, cursor) & inputchr & Mid(inputline, cursor + 1)
cursor += 1
ElseIf inputchr = Chr(08) And Cursor > 0 Then 'BkSp
cursor -= 1
inputline = Left(inputline, cursor) & Mid(inputline, cursor + 2)
ElseIf inputchr = Chr(255) & "S" And Cursor < Len(inputline) Then 'Del
inputline = Left(inputline, cursor) & Mid(inputline, cursor + 2)
ElseIf inputchr = Chr(255) + "M" And Cursor < Len(inputline) Then 'Right
Cursor += 1
ElseIf inputchr = Chr(255) + "K" And Cursor > 0 Then 'Left
Cursor -= 1
End If
If inputchr = Chr(27) Then 'Esc
Locate row, cursor0
Print Space(Len(inputline) + 1);
inputline = ""
cursor = 0
End If
Locate row, cursor0
Print Left(inputline, cursor) & Chr(95) & Mid(inputline, cursor + 1) & " ";
End If
Locate r, c
MutexUnlock(mutex)
Sleep sleeptime, 1
Loop Until inputchr = Chr(13)

If blank <> 0 Then
MutexLock(mutex)
r = CsrLin()
c = Pos()
Locate row, cursor0
Print Space(Len(inputline) + 1);
Locate r, c
MutexUnlock(mutex)
End If

Return inputline
End Function

From the example 1 on the Critical Sections page "Asynchronous method example using a mutex for all threads", now the running multi-threading code is waiting for the "quit" command in order to exit the program:

'   User thread algorithm:
'
'     Do
'     |  Mutexlock
'     |  .....
'     |  Critical section of code
'     |  .....
'     |  Mutexunlock
'     |  Sleep my_tempo, 1
'     Loop Until quit = true
'
'   There is no any advantage or disadvantage between threads for running their critical sections.

Function threadInput (ByVal row As Integer, ByVal column As Integer, ByRef prompt As String = "", _
ByVal sleeptime As Integer = 15, ByVal blank As Integer = 0, ByVal mutex As Any Ptr = 0 _
) As String
Dim As String inputchr
Dim As String inputline
Dim As Integer cursor
Dim As Integer cursor0
Dim As Integer r
Dim As Integer c

MutexLock(mutex)
r = CsrLin()
c = Pos()
Locate row, column
Print prompt & " _";
cursor0 = Pos() - 1
Locate r, c
MutexUnlock(mutex)

Do
MutexLock(mutex)
r = CsrLin()
c = Pos()
inputchr = Inkey
If inputchr <> "" Then
If inputchr >= Chr(32) And inputchr < Chr(255) Then
inputline = Left(inputline, cursor) & inputchr & Mid(inputline, cursor + 1)
cursor += 1
ElseIf inputchr = Chr(08) And Cursor > 0 Then 'BkSp
cursor -= 1
inputline = Left(inputline, cursor) & Mid(inputline, cursor + 2)
ElseIf inputchr = Chr(255) & "S" And Cursor < Len(inputline) Then 'Del
inputline = Left(inputline, cursor) & Mid(inputline, cursor + 2)
ElseIf inputchr = Chr(255) + "M" And Cursor < Len(inputline) Then 'Right
Cursor += 1
ElseIf inputchr = Chr(255) + "K" And Cursor > 0 Then 'Left
Cursor -= 1
End If
If inputchr = Chr(27) Then 'Esc
Locate row, cursor0
Print Space(Len(inputline) + 1);
inputline = ""
cursor = 0
End If
Locate row, cursor0
Print Left(inputline, cursor) & Chr(95) & Mid(inputline, cursor + 1) & " ";
End If
Locate r, c
MutexUnlock(mutex)
Sleep sleeptime, 1
Loop Until inputchr = Chr(13)

If blank <> 0 Then
MutexLock(mutex)
r = CsrLin()
c = Pos()
Locate row, cursor0
Print Space(Len(inputline) + 1);
Locate r, c
MutexUnlock(mutex)
End If

Return inputline
End Function

'------------------------------------------------------------------------------

Type UDT
Dim As Integer number
Dim As Integer tempo
Dim As Any Ptr pThread
Dim As ULongInt count
Static As Any Ptr pMutex
Static As Integer numberMax
Static As Integer quit
End Type
Dim As Any Ptr UDT.pMutex
Dim As Integer UDT.numberMax
Dim As Integer UDT.quit

Sub Counter (ByVal pt As UDT Ptr)
With *pt
Locate .number, .number, 0
Sleep 5, 1
.count += 1
Print .count;
End With
End Sub

Sub Thread (ByVal p As Any Ptr)
Dim As Integer myquit
Dim As UDT Ptr pUDT = p
With *pUDT
Do
MutexLock(.pMutex)
Counter(pUDT)
myquit = .quit
MutexUnlock(.pMutex)
Sleep .tempo, 1
Loop Until myquit = 1
End With
End Sub

Screen 12
UDT.numberMax = 6

Dim As UDT u(0 To UDT.numberMax)
For I As Integer = 0 To UDT.numberMax
u(I).number = i
u(I).tempo = 100 + 15 * I - 95 * Sgn(I)
Next I
UDT.pMutex = MutexCreate

Dim As Single t = Timer
For I As Integer = 1 To UDT.numberMax
u(I).pThread = ThreadCreate(@Thread, @u(I))
Next I

Do
Loop Until LCase(threadInput(8, 1, """quit"" for exit?", 10, 1, UDT.pMutex)) = "quit"

UDT.quit = 1

For I As Integer = 1 To UDT.numberMax
ThreadWait(u(I).pThread)
Next I
t = Timer - t

MutexDestroy(UDT.pMutex)
Dim As ULongInt c
For I As Integer = 1 To UDT.numberMax
c += u(I).count
Next I
Locate UDT.numberMax + 4, 1
Print CULngInt(c / t) & " increments per second"

Sleep

Note:

Otherwise, by using only graphics keywords (using the only position of the graphic cursor) as Line, Draw String, Put in the thread, induces a thread-safe procedure that is compatible with the Line Input keyword in the main code with no mutex:

Type UDT
Dim As Integer number
Dim As Integer tempo
Dim As Any Ptr pThread
Dim As ULongInt count
Dim As Any Ptr img
Static As Integer numberMax
Static As Integer quit
End Type
Dim As Integer UDT.numberMax
Dim As Integer UDT.quit

Const As String prompt = "Enter ""quit"" for exit"
Dim As String s

Sub Counter (ByVal pt As UDT Ptr) ' for a graphic character size 8x16
With *pt
Line .img, (0, 0)-(20 * 8 - 1, 16 - 1), 0, BF ' clearing the image buffer
Sleep 5, 1
.count += 1
Draw String .img, (0, 0), Str(.count) ' drawing in the image buffer
Put ((.number - 1) * 8, (.number - 1) * 16), .img, PSet ' copying the image buffer to screen
End With
End Sub

Sub Thread (ByVal p As Any Ptr) ' for a graphic character size 8x16
Dim As UDT Ptr pUDT = p
With *pUDT
.img = ImageCreate(20 * 8, 16) ' using an image buffer to avoid flickering
Do
Counter(pUDT)
Sleep .tempo, 1
Loop Until .quit = 1
ImageDestroy .img ' destroying the image buffer
End With
End Sub

Screen 12
UDT.numberMax = 6

Dim As UDT u(0 To UDT.numberMax)
For I As Integer = 0 To UDT.numberMax
u(I).number = i
u(I).tempo = 100 + 15 * I - 95 * Sgn(I)
Next I

Dim As Single t = Timer
For I As Integer = 1 To UDT.numberMax
u(I).pThread = ThreadCreate(@Thread, @u(I))
Next I

Do
Locate 8, 1, 0
Line Input; prompt; s
Locate , Len(prompt) + 3
Print Space(Len(s));
Loop Until LCase(s) = "quit"
UDT.quit = 1

For I As Integer = 1 To UDT.numberMax
ThreadWait(u(I).pThread)
Next I
t = Timer - t

Dim As ULongInt c
For I As Integer = 1 To UDT.numberMax
c += u(I).count
Next I
Locate UDT.numberMax + 4, 1
Print CULngInt(c / t) & " increments per second"

Sleep

- Screenlock...Scrennunlock blocks are not compatible with multi-threading (otherwise, the program hangs). This is why a mutex block must be used around each such block to ensure the mutual exclusion.
- The input keywords (like for keyboard, mouse) cannot be safely run when the screen is locked, therefore a such keyword must be outside of any Screenlock...Screenunlock block, so outside any Screenlock...Screenunlock block in its own thread, and protected of all Screenlock...Screenunlock blocks of other threads by a mutex block. Therefore, Getkey and Input, the statements that wait for keypress or line input are unusable, but Inkey that does not wait can work.

By applying some rules scrupulously, one can use Screenlock/Screenunlock inside the threads.
Principle of coding for all threads including the main code (main thread):

Do
	' instructions without display (printing/drawing, ...) neither input (input/inkey/mouse getting, ...)
	MutexLock(m)
	Screenlock
	' instructions with only display (printing/drawing, ...)
	Screenunlock
	' instructions with only input without waiting (inkey/mouse getting, ...)
	MutexUnlock(m)
	Sleep tempo, 1
Loop Until condition

For example, it is mandatory to use one Mutexlock...Mutexunlock block around each Screenlock...Screenunlock block, and one other around the Inkey instruction which itself must always be outside of any Screenlock...Screenunlock bloc:

Type ThreadUDT
Dim handle As Any Ptr
Static sync As Any Ptr
Static quit As Byte
End Type
Dim ThreadUDT.sync As Any Ptr
Dim ThreadUDT.quit As Byte

Function ClockTime () As String
Return Time
End Function

Function Counter () As Integer
Static C As Integer
C = (C + 1) Mod 1000000
Return C
End Function

Sub ProcedureThread (ByVal param As Any Ptr)
With *Cast(ThreadUDT Ptr, param)
Do
MutexLock(.sync)
ScreenLock
Line (544, 0)-(639, 49), 0, BF 'clear the print area
Sleep 100, 1
Locate 2, 71
Print ClockTime();
ScreenUnlock
MutexUnlock(.sync)
Sleep 100, 1
Loop Until .quit = 1
End With
End Sub

Screen 12
Locate 30, 2
Print "<q/Q> : quit";

Dim TTptr As ThreadUDT Ptr = New ThreadUDT
ThreadUDT.sync = MutexCreate
TTptr->handle = ThreadCreate(@ProcedureThread, TTptr)

Dim As String s
Do
MutexLock(ThreadUDT.sync)
ScreenLock
Line (296, 208)-(376, 256), 0, BF 'clear the print area
Sleep 100, 1
Locate 15,40
Print Using "######"; Counter();
ScreenUnlock
s = Inkey
MutexUnlock(ThreadUDT.sync)
Sleep 100, 1
Loop Until LCase(s) = "q"

ThreadUDT.quit = 1
ThreadWait(TTptr->handle)
MutexDestroy(ThreadUDT.sync)
Delete TTptr

Note: The Sleep x, 1 keyword just after the 'clear the print area' lines is only here to highlight the flickering if no screen locking is used.

Instead of "screen locking" (see the above paragraph), "video paging (double buffering or page flipping)" can more simply be used with multi-threading, but be careful that many states in the gfxlib2 are thread-dependent like Screenset (and also View settings, graphic cursor position, graphic colors, ...).
Therefore, the setting for the working page and the visible page must always be controlled in each thread code which want to work with a multi-video page configuration.

Example for a double buffering method (at each step, each thread needs to update the working page and copy it to the visible page, from within a mutual exclusion mutex code block):

Type ThreadUDT
Dim handle As Any Ptr
Static sync As Any Ptr
Static quit As Byte
End Type
Dim ThreadUDT.sync As Any Ptr
Dim ThreadUDT.quit As Byte

Function ClockTime () As String
Return Time
End Function

Function Counter () As Integer
Static C As Integer
C = (C + 1) Mod 1000000
Return C
End Function

Sub ProcedureThread (ByVal param As Any Ptr)
ScreenSet 1, 0 '' setting to define in each thread
With *Cast(ThreadUDT Ptr, param)
Do
MutexLock(.sync)
Line (544, 0)-(639, 49), 0, BF '' clear the print area
Sleep 100, 1
Locate 2, 71
Print ClockTime();
ScreenCopy
MutexUnlock(.sync)
Sleep 100, 1
Loop Until .quit = 1
End With
End Sub

Screen 12, , 2
ScreenSet 1, 0 '' setting to define in each thread
Locate 30, 2
Print "<q/Q> : quit";
ScreenCopy

Dim TTptr As ThreadUDT Ptr = New ThreadUDT
ThreadUDT.sync = MutexCreate
TTptr->handle = ThreadCreate(@ProcedureThread, TTptr)

Dim s As String
Do
MutexLock(ThreadUDT.sync)
Line (296, 208)-(376, 256), 0, BF '' clear the print area
Sleep 100, 1
Locate 15,40
Print Using "######"; Counter();
ScreenCopy
s = Inkey
MutexUnlock(ThreadUDT.sync)
Sleep 100, 1
Loop Until LCase(s) = "q"

ThreadUDT.quit = 1
ThreadWait(TTptr->handle)
MutexDestroy(ThreadUDT.sync)
Delete TTptr

Note: The Sleep x, 1 keyword just after the 'clear the print area' lines is only here to highlight the flickering if no double buffering is used.

Example for a two page flipping method (at each step, each thread needs to update and flip, from within the same exclusion mutex code block, the two screen pages):

Type ThreadUDT
Dim handle As Any Ptr
Static sync As Any Ptr
Static quit As Byte
End Type
Dim ThreadUDT.sync As Any Ptr
Dim ThreadUDT.quit As Byte

Function ClockTime () As String
Return Time
End Function

Function Counter () As Integer
Static C As Integer
C = (C + 1) Mod 1000000
Return C
End Function

Sub ProcedureThread (ByVal param As Any Ptr)
Dim p0 As Integer = 0
Dim p1 As Integer = 1
ScreenSet 1, 0 '' setting to define in each thread
With *Cast(ThreadUDT Ptr, param)
Do
MutexLock(.sync)
Dim s As String = ClockTime()
For I As Integer = 1 To 2 '' updating the two screen pages
Line (544, 0)-(639, 49), 0, BF '' clear the print area
Sleep 100, 1
Locate 2, 71
Print s;
ScreenSet p0, p1
Swap p0, p1
Next I
MutexUnlock(.sync)
Sleep 100, 1
Loop Until .quit = 1
End With
End Sub

Screen 12, , 2
Dim p0 As Integer = 0
Dim p1 As Integer = 1
ScreenSet 1, 0 '' setting to define in each thread
For I As Integer = 1 To 2 '' updating the two screen pages
Locate 30, 2
Print "<q/Q> : quit";
ScreenSet p0, p1
Swap p0, p1
Next I

Dim TTptr As ThreadUDT Ptr = New ThreadUDT
ThreadUDT.sync = MutexCreate
TTptr->handle = ThreadCreate(@ProcedureThread, TTptr)

Dim s As String
Do
MutexLock(ThreadUDT.sync)
Dim C As Integer = Counter()
For I As Integer = 1 To 2 '' updating the two screen pages
Line (296, 208)-(376, 256), 0, BF '' clear the print area
Sleep 100, 1
Locate 15,40
Print Using "######"; c;
ScreenSet p0, p1
Swap p0, p1
Next I
s = Inkey
MutexUnlock(ThreadUDT.sync)
Sleep 100, 1
Loop Until LCase(s) = "q"

ThreadUDT.quit = 1
ThreadWait(TTptr->handle)
MutexDestroy(ThreadUDT.sync)
Delete TTptr

Note: The Sleep x, 1 keyword just after the 'clear the print area' lines is only here to highlight the flickering if no two page flipping is used.

Note: In these two examples, a mutual exclusion mutex code block is mandatory in the two threads, not only because of using console statements + Inkey, but around also the graphics statements + Screencopy only because of using double buffering method (without anti-flickering process, the graphics statements could be outside the exclusion mutex code block).

The source code of gfxlib2 uses TLS (Thread Local Storage) to store many states, so many things are thread-specific.
Since gfxlib2 is thread-safe, mutual mutex exclusion between threads is not necessary for the graphics statements themselves (including Draw String).
In contrast, console statements such as Locate, Print, ... are not thread-safe as previously mentioned (for example, text cursor position is common to all threads).

Simple example showing that graphic states (such as graphic cursor position, graphic colors) are thread-dependent:

Screen 12

Sub thread(ByVal p As Any Ptr)
Color 10
PSet(150, 10)
For I As Integer = 1 To 40
Line -Step(10, 10)
Sleep 150, 1
Next I
Draw String Step (-40, 10), "user thread"
End Sub

Dim As Any Ptr p = ThreadCreate(@thread)

Color 14
PSet(10, 100)
For I As Integer = 1 To 24
Line -Step(10, 10)
Sleep 250, 1
Next I
Draw String Step (-40, 10), "main thread"

ThreadWait(p)

Color 15
Locate 4, 2
Print "Any key for exit"

Sleep

Example showing that graphics statements (such as Line and Draw String and Screencopy) in a thread can compete with console statements (such as Inkey) in another thread, without using any exclusion (by mutex):

#include "vbcompat.bi"

Screen 12, , 2
ScreenSet 1, 0
Color 0, 7
Cls

Dim Shared terminate As Integer = 0

Sub thread (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
Line (16, 432)-Step(96, 32), 11, BF 'clear print area
Sleep 100, 1
Draw String (24, 432), Format(Now,"dd/mm/yyyy"), 0
Draw String (32, 448), Format(Now,"hh:mm:ss"), 0
ScreenCopy
Sleep 100, 1
Loop Until terminate = 1
End Sub

Dim As String reply
Locate 2, 2
Print "Enter ""q"" to quit"
ScreenCopy

Dim p As Any Ptr = ThreadCreate(@thread)

Do
reply = Inkey
Sleep 100, 1
Loop Until LCase(reply) = "q"

Print " Stop the thread"
ScreenCopy
terminate=1
ThreadWait (p)
Print " Thread terminated"
ScreenCopy

Sleep

Note: The Sleep x, 1 keyword just after the 'clear the print area' line is only here to highlight the flickering if no double buffering is used.

From the above example, if the date displaying and the time displaying are now two separate threads, a mutual exclusion mutex code block between these two threads is mandatory, not due to the graphics statements themselves competing, but only due to the double buffering method used (against flickering) that puts competing these two threads:

#include "vbcompat.bi"

Screen 12, , 2
ScreenSet 1, 0
Color 0, 7
Cls

Dim Shared terminate As Integer = 0
Dim Shared mutex As Any Ptr

Sub thread1 (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
MutexLock(mutex)
Line (16, 432)-Step(96, 16), 11, BF 'clear the print area
Sleep 200, 1
Draw String (24, 432), Format(Now,"dd/mm/yyyy"), 0
ScreenCopy
MutexUnlock(mutex)
Sleep 100, 1
Loop Until terminate = 1
End Sub

Sub thread2 (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
MutexLock(mutex)
Line (16, 448)-Step(96, 16), 11, BF 'clear the print area
Sleep 100, 1
Draw String (32, 448), Format(Now,"hh:mm:ss"), 0
ScreenCopy
MutexUnlock(mutex)
Sleep 100, 1
Loop Until terminate = 1
End Sub

Dim As String reply
Locate 2, 2
Print "Enter ""q"" to quit"
ScreenCopy

mutex = MutexCreate
Dim p1 As Any Ptr = ThreadCreate(@thread1)
Dim p2 As Any Ptr = ThreadCreate(@thread2)

Do
reply = Inkey
Sleep 100, 1
Loop Until LCase(reply) = "q"

Print " Stop the threads"
ScreenCopy
terminate=1
ThreadWait (p1)
ThreadWait (p2)
MutexDestroy(mutex)
Print " Threads terminated"
ScreenCopy

Sleep

Note: The Sleep x, 1 keyword just after the 'clear the print area' lines is only here to highlight the flickering if no double buffering is used, or if no mutex is used.

Console statements (such as Locate, Print, Color, ...), as well as Locate and Print on Graphics window (but not Color on Graphics Window), and keyboard inputs (such as Inkey, Getkey, Input, ...) are not thread-safe:

- Thus when they are used in competing sections of different threads, mutual exclusion is mandatory by means of mutex locking blocks in which in addition code can restore states (such as text cursor position, console color, ...) at end of the block (after its own usage), as they were before (at begin of the block).
- But the Getkey or Input keyword cannot be enclosed inside a mutex locking block (as it can be do with the Inkey keyword), because as long as the keyboard input is not completed, the other threads in compete would be fully blocked (waiting for the mutex unlocking).

Example showing that the keywords Locate and Print are not thread-safe both when applied on a console window or when applied on a graphics window (the text cursor states being not thread dependent in the two cases):

Sub Thread (ByVal p As Any Ptr)
Locate Cast(Integer, p), Cast(Integer, p)
For I As Integer = 1 To 50 - 2 * Cast(Integer, p)
Sleep 20 * Cast(Integer, p), 1
Print Str(Cast(Integer, p));
Next I
End Sub

Sub test ()
Dim As Any Ptr p(1 To 9)
For I As Integer = 1 To 9
p(I) = ThreadCreate(@Thread, Cast(Any Ptr, I))
Sleep 25, 1
Next I
For I As Integer = 1 To 9
ThreadWait(p(I))
Next I
End Sub

Screen 0
test()
Locate 15, 1
Print "Any key to continue"
Sleep

Screen 12
test()
Locate 15, 1
Print "Any key to quit"
Sleep

Note: One can see that each thread does not write on its own line corresponding to its thread number (id between 1 and 9), on the console window and on the graphics window.

From the above example, the thread code has been completed in its competing sections by mutex locking blocks and by saving/restoring cursor states before/after its own cursor moving:

Dim Shared As Any Ptr mutex

Sub Thread (ByVal p As Any Ptr)
MutexLock(mutex)
Dim As Long l0 = Locate()
Locate Cast(Integer, p), Cast(Integer, p)
Dim As Long l = Locate()
Locate HiByte(LoWord(l0)), LoByte(LoWord(l0)), HiWord(l0)
MutexUnlock(mutex)
For I As Integer = 1 To 50 - 2 * Cast(Integer, p)
Sleep 20 * Cast(Integer, p), 1
MutexLock(mutex)
l0 = Locate()
Locate HiByte(LoWord(l)), LoByte(LoWord(l)), HiWord(l)
Print Str(Cast(Integer, p));
l = Locate()
Locate HiByte(LoWord(l0)), LoByte(LoWord(l0)), HiWord(l0)
MutexUnlock(mutex)
Next I
End Sub

Sub test ()
Dim As Any Ptr p(1 To 9)
For I As Integer = 1 To 9
p(I) = ThreadCreate(@Thread, Cast(Any Ptr, I))
Sleep 25, 1
Next I
For I As Integer = 1 To 9
ThreadWait(p(I))
Next I
End Sub

mutex = MutexCreate

Screen 0
test()
Locate 15, 1
Print "Any key to continue"
Sleep

Screen 12
test()
Locate 15, 1
Print "Any key to quit"
Sleep

MutexDestroy(mutex)

Note: One can see that each thread writes now on its own line corresponding to its thread number (id between 1 and 9), on the console window and on the graphics window.

Example showing that the Color keyword is not thread-safe when applied on a console window, but is thread-safe when applied on a graphics window (the color states being thread dependent in that case):

Sub Thread (ByVal p As Any Ptr)
Color Cast(Integer, p) + 8, Cast(Integer, p)
For I As Integer = 1 To 50 - 2 * Cast(Integer, p)
Print " " & Cast(Integer, p) & " ";
Sleep 20 * Cast(Integer, p), 1
Next I
End Sub

Sub test ()
Dim As Any Ptr p(1 To 9)
Locate 1, 1
For I As Integer = 1 To 9
p(I) = ThreadCreate(@Thread, Cast(Any Ptr, I))
Sleep 25, 1
Next I
For I As Integer = 1 To 9
ThreadWait(p(I))
Next I
Locate 16, 1
End Sub

Screen 0
test()
Print "Any key to continue"
Sleep

Screen 12
test()
Print "Any key to quit"
Sleep

Note: One can see that the foreground/background colors are not specific to the thread number (id between 1 and 9) on the console window, but this works great on the graphics window.

From the above example, the thread code has been completed in its competing sections by mutex locking blocks and by saving/restoring color states before/after its own color values usage:

Dim Shared As Any Ptr mutex

Sub Thread (ByVal p As Any Ptr)
MutexLock(mutex)
Dim As Ulong c0 = Color(Cast(Integer, p) + 8, Cast(Integer, p))
Dim As Ulong c = Color()
Color(LoWord(c0), HiWord(c0))
MutexUnlock(mutex)
For I As Integer = 1 To 50 - 2 * Cast(Integer, p)
MutexLock(mutex)
c0 = Color(LoWord(c), HiWord(c))
Print " " & Cast(Integer, p) & " ";
Color(LoWord(c0), HiWord(c0))
MutexUnlock(mutex)
Sleep 20 * Cast(Integer, p), 1
Next I
End Sub

Sub test ()
Dim As Any Ptr p(1 To 9)
Locate 1, 1
For I As Integer = 1 To 9
p(I) = ThreadCreate(@Thread, Cast(Any Ptr, I))
Sleep 25, 1
Next I
For I As Integer = 1 To 9
ThreadWait(p(I))
Next I
Locate 16, 1
End Sub

mutex = MutexCreate

Screen 0
test()
Print "Any key to continue"
Sleep

Screen 12
test()
Print "Any key to quit"
Sleep

MutexDestroy(mutex)

Note: One can see that the foreground/background colors are now specific to the thread number (id between 1 and 9) on the console window (obviously this always works on the graphics window).

Therefore, for using Getkey or Input in competing sections of threads:

- Only a single thread (for example, the main thread) can uses Getkey or Input in addition to console statements (such as Locate, Print, Color, ...) and also Inkey, in its competing sections.
- The other threads must not to use in their competing sections any console statement neither any keyboard input keyword, but can use by cons graphics statements (such as Pset, Line, Circle, Draw String, graphic Color, ...) which are themselves thread-safe (they can interlace graphically with the main thread without any problem).
- Input and Getkey also exclude the screen locking usage in competing sections of threads (double buffering is recommended as anti-flickering method).

Example showing that graphics statements (such as Line and Draw String and Screencopy) in a thread (user thread here) can compete with console statements (such as Locate and Print and Input) in another thread (main thread here), without using any mutual exclusion (by mutex):

#include "vbcompat.bi"

Screen 12, , 2
ScreenSet 1, 0
Color 0, 7
Cls

Dim Shared terminate As Integer = 0

Sub thread (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
Line (16, 432)-Step(96, 32), 11, BF 'clear the print area
Sleep 100, 1
Draw String (24, 432), Format(Now,"dd/mm/yyyy"), 0
Draw String (32, 448), Format(Now,"hh:mm:ss"), 0
ScreenCopy
Sleep 100, 1
Loop Until terminate = 1
End Sub

Dim As String reply
Locate 2, 2
Print "Enter ""quit"" to quit"
ScreenCopy

Dim p As Any Ptr = ThreadCreate(@thread)

Do
Locate 3, 2
Print Space(Len(reply) + 2);
Locate 3, 2
Input reply
Loop Until LCase(reply) = "quit"

Print " Stop the thread"
ScreenCopy
terminate=1
ThreadWait (p)
Print " Thread terminated"
ScreenCopy

Sleep

Note: The Sleep x, 1 keyword just after the 'clear the print area' line is only here to highlight the flickering if no double buffering is used (screen locking being forbidden by Input usage).

From the above example, if the date displaying and the time displaying are now two separate user threads, a mutual exclusion mutex code block between these two threads only is mandatory, not due to the graphics statements themselves competing, but only due to the double buffering method used (against flickering) that puts competing these two user threads only:

#include "vbcompat.bi"

Screen 12, , 2
ScreenSet 1, 0
Color 0, 7
Cls

Dim Shared terminate As Integer = 0
Dim Shared mutex As Any Ptr

Sub thread1 (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
MutexLock(mutex)
Line (16, 432)-Step(96, 16), 11, BF 'clear the print area
Sleep 200, 1
Draw String (24, 432), Format(Now,"dd/mm/yyyy"), 0
ScreenCopy
MutexUnlock(mutex)
Sleep 100, 1
Loop Until terminate = 1
End Sub

Sub thread2 (ByVal param As Any Ptr)
ScreenSet 1, 0
Do
MutexLock(mutex)
Line (16, 448)-Step(96, 16), 11, BF 'clear the print area
Sleep 100, 1
Draw String (32, 448), Format(Now,"hh:mm:ss"), 0
ScreenCopy
MutexUnlock(mutex)
Sleep 100, 1
Loop Until terminate = 1
End Sub

Dim As String reply
Locate 2, 2
Print "Enter ""quit"" to quit"
ScreenCopy

mutex = MutexCreate
Dim p1 As Any Ptr = ThreadCreate(@thread1)
Dim p2 As Any Ptr = ThreadCreate(@thread2)

Do
Locate 3, 2
Print Space(Len(reply) + 2);
Locate 3, 2
Input reply
Loop Until LCase(reply) = "quit"

Print " Stop the threads"
ScreenCopy
terminate=1
ThreadWait (p1)
ThreadWait (p2)
MutexDestroy(mutex)
Print " Threads terminated"
ScreenCopy

Sleep

Note: The Sleep x, 1 keyword just after the 'clear the print area' lines is only here to highlight the flickering if no double buffering is used (screen locking being forbidden by Input usage).

There is still some doubt about the perfect behavior of the keyword Sleep in a multi-threading context.

It is therefore advisable to take the following precautions for its use:

- If it is absolutely necessary in a critical section of a thread, the syntax Sleep x or Sleep x, 0, because inducing an internal test of a key-press, must for greatest safety be preferably treated in the same way as the Inkey keyword to avoid as much as possible any concurrent conflict with other threads.
- Otherwise, the syntax Sleep x, 1 (inducing no internal test of key-press) is rather advised when there is no protection by mutual exclusion, which is very often the case in order to release time-slice for the other threads.

A simple 'threadUDT' base Class can be defined as follows:

- with a private 'Any Ptr' non-static member field for each handle,
- with a private 'Any Ptr' static member field for one shared mutex,
- with a private 'Any Ptr' static member field for one shared conditional variable,
- with its own public member procedures 'Sub()' calling the corresponding built-in procedures for multi-threading (with same procedure names), including also value integrity tests on the 3 above pointers (non-static procedures for the 3 'thread...()' member Subs, and static procedures for the 4 'mutex...()' member Subs and the 5 'cond...()' member Subs),
- with an abstract private 'Sub()' thread to be overridden by another 'Sub()' from inside the user derived Type (therefore its static address is available in the virtual table of the object, and the hidden 'This' parameter passed by reference is compatible with the 'Any Ptr' parameter to be passed to the thread).

#include once "fbthread.bi"
Type threadUDT Extends Object
Public:
Declare Sub ThreadCreate ()
Declare Sub ThreadWait ()
Declare Sub threadDetach ()
Declare Static Sub MutexCreate ()
Declare Static Sub MutexLock ()
Declare Static Sub MutexUnlock ()
Declare Static Sub MutexDestroy ()
Declare Static Sub CondCreate ()
Declare Static Sub CondWait ()
Declare Static Sub CondSignal ()
Declare Static Sub CondBroadcast ()
Declare Static Sub CondDestroy ()
Private:
Declare Abstract Sub thread ()
Dim As Any Ptr pThread
Static As Any Ptr pMutex
Static As Any Ptr pCond
End Type
Dim As Any Ptr threadUDT.pMutex
Dim As Any Ptr threadUDT.pCond

Sub threadUDT.ThreadCreate ()
If This.pThread = 0 Then
This.pThread = .ThreadCreate(Cast(Any Ptr Ptr Ptr, @This)[0][0], @This)
End If
End Sub

Sub threadUDT.ThreadWait ()
If This.pThread > 0 Then
.ThreadWait(This.pThread)
This.pThread = 0
End If
End Sub

Sub threadUDT.threadDetach ()
If This.pThread > 0 Then
.Threaddetach(This.pThread)
This.pThread = 0
End If
End Sub

Sub threadUDT.MutexCreate ()
If threadUDT.pMutex = 0 Then
threadUDT.pMutex = .MutexCreate
End If
End Sub

Sub threadUDT.MutexLock ()
If threadUDT.pMutex > 0 Then
.MutexLock(threadUDT.pMutex)
End If
End Sub

Sub threadUDT.MutexUnlock ()
If threadUDT.pMutex > 0 Then
.MutexUnlock(threadUDT.pMutex)
End If
End Sub

Sub threadUDT.MutexDestroy ()
If threadUDT.pMutex > 0 Then
.MutexDestroy(threadUDT.pMutex)
threadUDT.pMutex = 0
End If
End Sub

Sub threadUDT.CondCreate ()
If threadUDT.pCond = 0 Then
threadUDT.pCond = .CondCreate
End If
End Sub

Sub threadUDT.CondWait ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondWait(threadUDT.pCond, threadUDT.pMutex)
End If
End Sub

Sub threadUDT.CondSignal ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondSignal(threadUDT.pCond)
End If
End Sub

Sub threadUDT.CondBroadcast ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondBroadcast(threadUDT.pCond)
End If
End Sub

Sub threadUDT.CondDestroy ()
If threadUDT.pCond > 0 Then
.CondDestroy(threadUDT.pCond)
threadUDT.pCond = 0
End If
End Sub

From the example 2 on the Critical Sections page "Synchronous method example using a condwait then a condbroadcast (and a mutex) for all threads", now the user implementation is modified to be compatible with the base Class 'threadUDT':

#include once "fbthread.bi"
Type threadUDT Extends Object
Public:
Declare Sub ThreadCreate ()
Declare Sub ThreadWait ()
Declare Sub threadDetach ()
Declare Static Sub MutexCreate ()
Declare Static Sub MutexLock ()
Declare Static Sub MutexUnlock ()
Declare Static Sub MutexDestroy ()
Declare Static Sub CondCreate ()
Declare Static Sub CondWait ()
Declare Static Sub CondSignal ()
Declare Static Sub CondBroadcast ()
Declare Static Sub CondDestroy ()
Private:
Declare Abstract Sub thread ()
Dim As Any Ptr pThread
Static As Any Ptr pMutex
Static As Any Ptr pCond
End Type
Dim As Any Ptr threadUDT.pMutex
Dim As Any Ptr threadUDT.pCond

Sub threadUDT.ThreadCreate ()
If This.pThread = 0 Then
This.pThread = .ThreadCreate(Cast(Any Ptr Ptr Ptr, @This)[0][0], @This)
End If
End Sub

Sub threadUDT.ThreadWait ()
If This.pThread > 0 Then
.ThreadWait(This.pThread)
This.pThread = 0
End If
End Sub

Sub threadUDT.threadDetach ()
If This.pThread > 0 Then
.Threaddetach(This.pThread)
This.pThread = 0
End If
End Sub

Sub threadUDT.MutexCreate ()
If threadUDT.pMutex = 0 Then
threadUDT.pMutex = .MutexCreate
End If
End Sub

Sub threadUDT.MutexLock ()
If threadUDT.pMutex > 0 Then
.MutexLock(threadUDT.pMutex)
End If
End Sub

Sub threadUDT.MutexUnlock ()
If threadUDT.pMutex > 0 Then
.MutexUnlock(threadUDT.pMutex)
End If
End Sub

Sub threadUDT.MutexDestroy ()
If threadUDT.pMutex > 0 Then
.MutexDestroy(threadUDT.pMutex)
threadUDT.pMutex = 0
End If
End Sub

Sub threadUDT.CondCreate ()
If threadUDT.pCond = 0 Then
threadUDT.pCond = .CondCreate
End If
End Sub

Sub threadUDT.CondWait ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondWait(threadUDT.pCond, threadUDT.pMutex)
End If
End Sub

Sub threadUDT.CondSignal ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondSignal(threadUDT.pCond)
End If
End Sub

Sub threadUDT.CondBroadcast ()
If threadUDT.pCond > 0 And threadUDT.pMutex > 0 Then
.CondBroadcast(threadUDT.pCond)
End If
End Sub

Sub threadUDT.CondDestroy ()
If threadUDT.pCond > 0 Then
.CondDestroy(threadUDT.pCond)
threadUDT.pCond = 0
End If
End Sub

'------------------------------------------------------------------------------

Type UDT Extends threadUDT
Declare Sub counter ()
Declare Sub thread ()
Dim As Integer number
Dim As Integer tempo
Dim As ULongInt count
Static As Integer threadPriorityNumber
Static As Integer numberMax
Static As Integer quit
End Type
Dim As Integer UDT.threadPriorityNumber
Dim As Integer UDT.numberMax
Dim As Integer UDT.quit

Sub UDT.counter ()
Locate This.number, This.number, 0
Sleep 5, 1
This.count += 1
Print This.count;
Locate This.number, 30 + This.number, 0
End Sub

Sub UDT.Thread ()
Dim As Integer myquit
Do
This.MutexLock()
While UDT.threadPriorityNumber <> This.number '' synchronous condwait for expected condition
This.CondWait()
Wend
This.counter()
myquit = UDT.quit
UDT.threadPriorityNumber = (UDT.threadPriorityNumber + 1) Mod (UDT.numberMax + 1)
This.CondBroadcast()
This.MutexUnlock()
Sleep This.tempo, 1
Loop Until myquit = 1
End Sub

UDT.numberMax = 6
Dim As UDT u(0 To UDT.numberMax)
For I As Integer = 0 To UDT.numberMax
u(I).number = i
u(I).tempo = 100 + 15 * I - 95 * Sgn(I)
Next I
UDT.MutexCreate()
UDT.CondCreate()

Dim As Single t = Timer
For I As Integer = 1 To UDT.numberMax
u(I).ThreadCreate()
Next I

Dim As String s
Do
UDT.MutexLock()
While UDT.threadPriorityNumber <> u(0).number
UDT.CondWait()
Wend
s = Inkey
If s <> "" Then
UDT.quit = 1
End If
UDT.threadPriorityNumber = (UDT.threadPriorityNumber + 1) Mod (UDT.numberMax + 1)
UDT.CondBroadcast()
UDT.MutexUnlock()
Sleep u(0).tempo, 1
Loop Until s <> ""

For I As Integer = 1 To UDT.numberMax
u(I).ThreadWait()
Next I
t = Timer - t

UDT.MutexDestroy()
UDT.CondDestroy()
Dim As ULongInt c
For I As Integer = 1 To UDT.numberMax
c += u(I).count
Next I
Locate UDT.numberMax+2, 1
Print CULngInt(c / t) & " increments per second"

Sleep

One might think that the first code line of the thread is always executed at least after the 'ThreadCreate()' returns, but this is neither guaranteed nor even observed.

One can estimate the delay (positive or negative) between the 'ThreadCreate()' return and the the thread start, by a time memorization as similar as possible between the line following 'ThreadCreate()' and the first thread code line (the delay calculation is executed after the end of the thread).
After a while of observation, one can find both small negative values and large positive values.

Interesting to see the min time, average time, and max time, between the executing start of thread body and the returning point of 'ThreadCreate()':

Dim As Any Ptr ptid
Dim As Double t0
Dim As Any Ptr p0 = @t0
Dim As Double t1
Dim As Double count
Dim As Single tmean
Dim As Single tmin = 10 '' start value
Dim As Single tmax = -10 '' start value

Sub myThread (ByVal p As Any Ptr)
*Cast(Double Ptr, p) = Timer '' similar code line as in main code
End Sub

Print "Tmin/Tmean/Tmax between begin of thread code and return from ThreadCreate() :"
Do
count += 1
ptid = ThreadCreate(@myThread, @t1)
*Cast(Double Ptr, p0) = Timer '' similar code line as in thread code

ThreadWait(ptid)

tmean = (tmean * (count - 1) + (t1 - t0)) / count
If t1 - t0 < tmin Or t1 - t0 > tmax Then
If t1 - t0 < tmin Then
tmin = t1 - t0
End If
If t1 - t0 > tmax Then
tmax = t1 - t0
End If
Print Time; Using " Tmin=+###.###### ms Tmean=+###.###### ms Tmax=+###.###### ms"; tmin * 1000; tmean * 1000; tmax * 1000
End If
Loop Until Inkey <> ""

Output (for example):

Tmin/Tmean/Tmax between begin of thread code and return from ThreadCreate() :
21:30:13    Tmin=  +0.151800 ms    Tmean=  +0.151800 ms    Tmax=  +0.151800 ms
21:30:13    Tmin=  +0.006000 ms    Tmean=  +0.078900 ms    Tmax=  +0.151800 ms
21:30:13    Tmin=  +0.006000 ms    Tmean=  +0.098394 ms    Tmax=  +0.172500 ms
21:30:13    Tmin=  +0.006000 ms    Tmean=  +0.121555 ms    Tmax=  +0.884900 ms
21:30:45    Tmin=  +0.006000 ms    Tmean=  +0.055810 ms    Tmax=  +1.104200 ms
21:30:54    Tmin=  +0.006000 ms    Tmean=  +0.055764 ms    Tmax=  +4.056600 ms
21:31:44    Tmin=  -0.116300 ms    Tmean=  +0.055516 ms    Tmax=  +4.056600 ms
21:32:10    Tmin=  -0.136800 ms    Tmean=  +0.057177 ms    Tmax=  +4.056600 ms
21:32:12    Tmin=  -0.150300 ms    Tmean=  +0.057265 ms    Tmax=  +4.056600 ms
21:33:17    Tmin=  -0.150300 ms    Tmean=  +0.060048 ms    Tmax=  +4.979900 ms
21:33:18    Tmin=  -0.150300 ms    Tmean=  +0.060157 ms    Tmax=  +7.086300 ms
21:33:23    Tmin=  -0.150600 ms    Tmean=  +0.060347 ms    Tmax=  +7.086300 ms
21:33:38    Tmin=  -0.205900 ms    Tmean=  +0.060878 ms    Tmax=  +7.086300 ms
21:35:30    Tmin=  -0.208700 ms    Tmean=  +0.061315 ms    Tmax=  +7.086300 ms

Note:

If the user safely wish to always delay the thread execution at least after some code lines following the 'ThreadCreate()' line, a mutual exclusion between the 'ThreadCreate()' line and the start of the thread body can be used as this principle follows:

Dim Shared As Any Ptr pMutexForThreadStart

'-------------------------------------------

Sub Thread (ByVal p As Any Ptr)
MutexLock(pMutexForThreadStart)
MutexUnlock(pMutexForThreadStart)
'
' user thread body
'
End Sub

'--------------------------------------------

'
' user main code
'
pMutexForThreadStart = MutexCreate()
'
' user main code continues
'
MutexLock(pMutexForThreadStart)
Dim As Any Ptr pThread = ThreadCreate(@Thread)
'
' lines of code to be executed before the executing start of the user body of the thread
'
MutexUnlock(pMutexForThreadStart)
'
' user main code continues
'
ThreadWait(pThread)
MutexDestroy(pMutexForThreadStart)

The synchronization waiting phase of each thread should not consume any CPU resources like 'Sleep', which is the case of 'MutexLock()' and 'CondWait()' instructions.
Thread synchronization by mutual exclusions or by conditional variables adds latency to the initial execution time of the threads, but this latency (a few microseconds) is infinitely shorter than that of a simple simple wait loop (a few milliseconds at best) containing the shortest sleep ('Sleep 1, 1') with a flag test.

The following code allows to estimate this synchronization latency between the main thread and a child thread, by using either simple flags, either mutual exclusions, or conditional variables:

Dim Shared As Any Ptr mutex0, mutex1, mutex2, mutex, cond1, cond2, pt
Dim Shared As Integer flag1, flag2
Dim As Double t

'----------------------------------------------------------------------------------

#if defined(__FB_WIN32__)
Declare Function _setTimer Lib "winmm" Alias "timeBeginPeriod"(ByVal As Ulong = 1) As Long
Declare Function _resetTimer Lib "winmm" Alias "timeEndPeriod"(ByVal As Ulong = 1) As Long
#endif

Sub ThreadFlag(ByVal p As Any Ptr)
MutexUnlock(mutex0) '' unlock mutex for main thread
For I As Integer = 1 To 100
While flag1 = 0
Sleep 1, 1
Wend
flag1 = 0
' only child thread code runs (location for example)
flag2 = 1
Next I
End Sub

mutex0 = MutexCreate()
MutexLock(mutex0)

pt = ThreadCreate(@ThreadFlag)
MutexLock(mutex0) '' wait for thread launch (mutex unlock from child thread)
Print "Thread synchronization latency by simple flags:"
#if defined(__FB_WIN32__)
_setTimer()
Print "(in high resolution OS cycle period)"
#else
Print "(in normal resolution OS cycle period)"
#endif
t = Timer
For I As Integer = 1 To 100
flag1 = 1
While flag2 = 0
Sleep 1, 1
Wend
flag2 = 0
' only main thread code runs (location for example)
Next I
t = Timer - t
#if defined(__FB_WIN32__)
_resetTimer()
#endif
ThreadWait(pt)
Print Using "####.## milliseconds per double synchronization (round trip)"; t * 10
Print

MutexDestroy(mutex0)

'----------------------------------------------------------------------------------

Sub ThreadMutex(ByVal p As Any Ptr)
MutexUnlock(mutex0) '' unlock mutex for main thread
For I As Integer = 1 To 100000
MutexLock(mutex1) '' wait for mutex unlock from main thread
' only child thread code runs
MutexUnlock(mutex2) '' unlock mutex for main thread
Next I
End Sub

mutex0 = MutexCreate()
mutex1 = MutexCreate()
mutex2 = MutexCreate()
MutexLock(mutex0)
MutexLock(mutex1)
MutexLock(mutex2)

pt = ThreadCreate(@ThreadMutex)
MutexLock(mutex0) '' wait for thread launch (mutex unlock from child thread)
Print "Thread synchronization latency by mutual exclusions:"
t = Timer
For I As Integer = 1 To 100000
MutexUnlock(mutex1) '' mutex unlock for child thread
MutexLock(mutex2) '' wait for mutex unlock from child thread
' only main thread code runs
Next I
t = Timer - t
ThreadWait(pt)
Print Using "####.## microseconds per double synchronization (round trip)"; t * 10
Print

MutexDestroy(mutex0)
MutexDestroy(mutex1)
MutexDestroy(mutex2)

'----------------------------------------------------------------------------------

Sub ThreadCondVar(ByVal p As Any Ptr)
MutexUnlock(mutex0) '' unlock mutex for main thread
For I As Integer = 1 To 100000
MutexLock(mutex)
While flag1 = 0
CondWait(cond1, mutex) '' wait for conditional signal from main thread
Wend
flag1 = 0
' only child thread code runs (location for example)
flag2 = 1
CondSignal(cond2) '' send conditional signal to main thread
MutexUnlock(mutex)
Next I
End Sub

mutex0 = MutexCreate()
mutex = MutexCreate()
MutexLock(mutex0)
cond1 = CondCreate()
cond2 = CondCreate()

pt = ThreadCreate(@ThreadCondVar)
MutexLock(mutex0) '' wait for thread launch (mutex unlock from child thread)
Print "Thread synchronization latency by conditional variables:"
t = Timer
For I As Integer = 1 To 100000
MutexLock(mutex)
flag1 = 1
CondSignal(cond1) '' send conditional signal to main thread
While flag2 = 0
CondWait(Cond2, mutex) '' wait for conditional signal from child thread
Wend
flag2 = 0
' only child thread code runs (location for example)
MutexUnlock(mutex)
Next I
t = Timer - t
ThreadWait(pt)
Print Using "####.## microseconds per double synchronization (round trip)"; t * 10
Print

MutexDestroy(mutex0)
MutexDestroy(mutex)
CondDestroy(cond1)
CondDestroy(cond2)

'----------------------------------------------------------------------------------

Sleep

Example of results:

Thread synchronization latency by simple flags:
(in high resolution OS cycle period)
   2.02 milliseconds per double synchronization (round trip)

Thread synchronization latency by mutual exclusions:
   5.93 microseconds per double synchronization (round trip)

Thread synchronization latency by conditional variables:
   7.54 microseconds per double synchronization (round trip)

Example of thread synchronization for executing in concurrent or exclusive mode by using conditional variable:

Dim Shared As Any Ptr pt, mutex1, mutex2, cond1, cond2
Dim Shared As Integer quit, flag1, flag2

Print "'1': Main thread procedure running (alone)"
Print "'2': Child thread procedure running (alone)"
Print "'-': Main thread procedure running (with the one of child thread)"
Print "'=': Child thread procedure running (with the one of main thread)"
Print

Sub Prnt(ByRef s As String, ByVal n As Integer)
For I As Integer = 1 To n
Print s;
Sleep 20, 1
Next I
End Sub

Sub ThreadCondCond(ByVal p As Any Ptr)
Do
MutexLock(mutex1)
While flag1 = 0 '' test flag set from main thread
CondWait(cond1, mutex1) '' wait for conditional signal from main thread
Wend
flag1 = 0 '' reset flag
MutexUnlock(mutex1)
If quit = 1 Then Exit Sub '' exit the threading loop
Prnt("=", 10)
MutexLock(mutex2)
flag2 = 1 '' set flag to main thread
CondSignal(cond2) '' send conditional signal to main thread
Prnt("2", 10)
MutexUnlock(mutex2)
Loop
End Sub

mutex1 = MutexCreate()
mutex2 = MutexCreate()
cond1 = CondCreate()
cond2 = CondCreate()

pt = ThreadCreate(@ThreadCondCond)
For I As Integer = 1 To 5
MutexLock(mutex1)
flag1 = 1 '' set flag to child thread
CondSignal(cond1) '' send conditional signal to child thread
MutexUnlock(mutex1)
Prnt("-", 10)
MutexLock(mutex2)
While flag2 = 0 '' test flag set from child thread
CondWait(Cond2, mutex2) '' wait for conditional signal from child thread
Wend
flag2 = 0 '' reset flag
Prnt("1", 10)
MutexUnlock(mutex2)
Next I

MutexLock(mutex1)
quit = 1 '' set quit for child thread
flag1 = 1
CondSignal(cond1) '' send conditional signal to child thread
MutexUnlock(mutex1)
ThreadWait(pt) '' wait for child thread to end
Print

MutexDestroy(mutex1)
MutexDestroy(mutex2)
CondDestroy(cond1)
CondDestroy(cond2)

Sleep

Output:

'1': Main thread procedure running (alone)
'2': Child thread procedure running (alone)
'-': Main thread procedure running (with the one of child thread)
'=': Child thread procedure running (with the one of main thread)

-==-=-=--==--==-=-=-22222222221111111111-=-=-=-==--==-=--==-22222222221111111111-=-=-==-=-=--=-=-==-22222222221111111111
-=-=-=-=-=-=--==-=-=22222222221111111111-==--==--==-=-=--==-22222222221111111111

The example below works with 6 threads (in addition to the main thread).
The first 3 threads (#1 to #3) are waiting on their own condition variable, while the last 3 threads (#4 to #6) are waiting on a same other condition variable.
These last 3 threads are awakened either on 'CondSignal()' or on 'CondBroadcast()'.

Type ThreadData
Dim As Integer id
Dim As Any Ptr mutex
Dim As Any Ptr cond
Dim As Boolean flag
Dim As Boolean quit
Dim As Any Ptr handle
Declare Static Sub Thread(ByVal p As Any Ptr)
End Type

Sub ThreadData.Thread(ByVal p As Any Ptr)
Dim As ThreadData Ptr pdata = p
Print " thread #" & pdata->id & " is running"
Do
MutexLock(pdata->mutex)
While pdata->flag = False
CondWait(pdata->cond, pdata->mutex)
Wend
pdata->flag = False
MutexUnlock(pdata->mutex)
If pdata->quit = False Then
Print " thread #" & pdata->id & " is signaled"
Else
Exit Do
End If
Loop
Print " thread #" & pdata->id & " is finishing"
End Sub

Dim As Any Ptr mutex = MutexCreate()
Dim As Any Ptr cond(0 To 3) = {CondCreate(), CondCreate(), CondCreate(), CondCreate()}
Dim As ThreadData mythreads(1 To 6) = {Type(1, mutex, cond(1)), Type(2, mutex, cond(2)), Type(3, mutex, cond(3)), _
Type(4, mutex, cond(0)), Type(5, mutex, cond(0)), Type(6, mutex, cond(0))}

Print "Threads from #1 to #6 are created:"
For I As Integer = LBound(mythreads) To UBound(mythreads)
mythreads(I).handle = ThreadCreate(@ThreadData.Thread, @mythreads(I))
Next I
Sleep 1000, 1 '' wait for all threads started
Print
Print "----------------------------------------------------------"
Print

For I As Integer = 3 To 1 Step -1
Print "Send a CondSignal to thread #" & I &":"
MutexLock(mutex)
mythreads(I).flag = True
CondSignal(cond(I))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for the thread loop completed
Print
Next I
Print "----------------------------------------------------------"
Print

Print "Send a single CondBroadcast to all threads from #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondBroadcast(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for all thread loops completed
Print "Send a single CondBroadcast to all threads from #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondBroadcast(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for all thread loops completed
Print "Send a single CondBroadcast to all threads from #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondBroadcast(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for all thread loops completed
Print
Print "----------------------------------------------------------"
Print

Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print

Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print

Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print "Send a CondSignal to any thread among #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
Next I
CondSignal(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for a thread loop completed
Print
Print "----------------------------------------------------------"
Print

For I As Integer = 1 To 3
Print "Send to finish a CondSignal to thread #" & I &":"
MutexLock(mutex)
mythreads(I).flag = True
mythreads(I).quit = True
CondSignal(cond(I))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for the thread loop completed
Print
Next I
Print "----------------------------------------------------------"
Print

Print "Send to finish a single CondBroadcast to all threads from #4 to #6:"
MutexLock(mutex)
For I As Integer = 4 To 6
mythreads(I).flag = True
mythreads(I).quit = True
Next I
CondBroadcast(cond(0))
MutexUnlock(mutex)
Sleep 1000, 1 '' wait for all thread loops completed
Print
Print "----------------------------------------------------------"
Print

For I As Integer = 1 To 3
ThreadWait(mythreads(I).handle)
CondDestroy(cond(I))
Next I
For I As Integer = 4 To 6
ThreadWait(mythreads(I).handle)
Next I
Print "All threads from #1 to #6 are finished."
Print

MutexDestroy(mutex)
CondDestroy(cond(0))

Sleep

Output (for example):

Threads from #1 to #6 are created:
   thread #1 is running
   thread #3 is running
   thread #2 is running
   thread #5 is running
   thread #4 is running
   thread #6 is running

----------------------------------------------------------

Send a CondSignal to thread #3:
   thread #3 is signaled

Send a CondSignal to thread #2:
   thread #2 is signaled

Send a CondSignal to thread #1:
   thread #1 is signaled

----------------------------------------------------------

Send a single CondBroadcast to all threads from #4 to #6:
   thread #5 is signaled
   thread #6 is signaled
   thread #4 is signaled
Send a single CondBroadcast to all threads from #4 to #6:
   thread #6 is signaled
   thread #4 is signaled
   thread #5 is signaled
Send a single CondBroadcast to all threads from #4 to #6:
   thread #5 is signaled
   thread #4 is signaled
   thread #6 is signaled

----------------------------------------------------------

Send a CondSignal to any thread among #4 to #6:
   thread #5 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #4 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #6 is signaled

Send a CondSignal to any thread among #4 to #6:
   thread #5 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #4 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #6 is signaled

Send a CondSignal to any thread among #4 to #6:
   thread #5 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #4 is signaled
Send a CondSignal to any thread among #4 to #6:
   thread #6 is signaled

----------------------------------------------------------

Send to finish a CondSignal to thread #1:
   thread #1 is finishing

Send to finish a CondSignal to thread #2:
   thread #2 is finishing

Send to finish a CondSignal to thread #3:
   thread #3 is finishing

----------------------------------------------------------

Send to finish a single CondBroadcast to all threads from #4 to #6:
   thread #4 is finishing
   thread #5 is finishing
   thread #6 is finishing

----------------------------------------------------------

All threads from #1 to #6 are finished.

The delay between the return of 'ThreadCreate()' and the start of the thread code (first line of thread code) can be estimated at about 50 microseconds on average, but can go up to a few milliseconds at worst.
(see FAQ 12. What is the execution delay of the code of a thread after the thread is created by 'ThreadCreate'?)

This is why a child thread can be launched only once (by a constructor for example) and execute a permanent waiting loop of user tasks (to avoid a thread launch latency each time), then at end stopped (by a destructor).
The synchronization between the main thread and the child thread (start of each user task and user task completed) can be managed by means of 2 mutexes.

Example to estimate the average time to execute a sequence of user tasks (with empty user procedure body):

- either launched by successive threads,
- or launched by a single thread.

Sub userTask(ByVal p As Any Ptr) '' task to execute
End Sub

Dim As Double t

'---------------------------------------------------------------------------

Print "Successive (empty) user tasks executed by one thread for each:"
t = Timer
For i As Integer = 1 To 10000
Dim As Any Ptr p = ThreadCreate(@userTask)
ThreadWait(p)
Next i
t = Timer - t
Print Using "######.### microdeconds per user task"; t * 100
Print

'---------------------------------------------------------------------------

Type thread
Public:
Dim As Sub(ByVal p As Any Ptr) task '' pointer to user task
Declare Sub Launch() '' launch user task
Declare Sub Wait() '' wait for user task completed
Declare Constructor()
Declare Destructor()
Private:
Dim As Any Ptr mutex1
Dim As Any Ptr mutex2
Dim As Any Ptr handle
Dim As Boolean quit
Declare Static Sub proc(ByVal pthread As thread Ptr)
End Type

Constructor thread()
This.mutex1 = MutexCreate
This.mutex2 = MutexCreate
MutexLock(This.mutex1)
MutexLock(This.mutex2)
This.handle = ThreadCreate(CPtr(Any Ptr, @thread.proc), @This)
End Constructor

Destructor thread()
This.quit = True
MutexUnlock(This.mutex1)
ThreadWait(This.handle)
MutexDestroy(This.mutex1)
MutexDestroy(This.mutex2)
End Destructor

Sub thread.proc(ByVal pthread As thread Ptr)
Do
MutexLock(pthread->mutex1) '' wait for launching task
If pthread->quit = True Then Exit Sub
pthread->task(pthread)
MutexUnlock(pthread->mutex2) '' task completed
Loop
End Sub

Sub thread.Launch()
MutexUnlock(This.mutex1)
End Sub

Sub thread.Wait()
MutexLock(This.mutex2)
End Sub

Print "Successive (empty) user tasks executed by a single thread for all:"
t = Timer
Dim As thread Ptr pThread = New Thread
pThread->task = @userTask
For i As Integer = 1 To 10000
pThread->Launch()
pThread->Wait()
Next i
Delete pThread
t = Timer - t
Print Using "######.### microdeconds per user task"; t * 100
Print

Sleep

Output (for example):

Successive (empty) user tasks executed by one thread for each:
   145.004 microdeconds per user task

Successive (empty) user tasks executed by a single thread for all:
	 6.691 microdeconds per user task

Each core has its own cache memory that allows to buffer the useful data (in read and write) of the shared memory.
Consequently, a cache coherence algorithm between cores is executed, to keep, in case of writing in the cache and for the common memory areas between caches, the most recent values among all the caches concerned.
It is this algorithm which penalizes the performance of multi-threading in the case of multiple accesses in shared memory, even more particularly in write mode.

It is therefore necessary to limit as much as possible the access of threads to shared memory, even more in writing.
For example, all intermediate results of threads could be performed in local memory, and only the final useful ones put in shared memory.

Example of a member thread procedures computing the sum of the first N integers, by accumulation directly in the shared memory ('SumUpTo_1()') or internally in its local memory before copy back ('SumUpTo_2'()'):

Type Thread
Dim As UInteger valueIN
Dim As Double valueOUT
Dim As Any Ptr pHandle
Declare Static Sub SumUpTo_1(ByVal pt As Thread Ptr)
Declare Static Sub SumUpTo_2(ByVal pt As Thread Ptr)
End Type

Sub Thread.SumUpTo_1(ByVal pt As Thread Ptr)
pt->valueOut = 0
For I As UInteger = 1 To pt->valueIN
pt->valueOUT += I
Next I
End Sub

Sub Thread.SumUpTo_2(ByVal pt As Thread Ptr)
Dim As Double value = 0
For I As UInteger = 1 To pt->valueIN
value += I
Next I
pt->valueOUT = value
End Sub

Sub MyThreads(ByVal pThread As Any Ptr, ByVal threadNB As UInteger = 1)
Dim As Thread td(1 To threadNB)
Dim As Double t

t = Timer
For i As Integer = 1 To threadNB
td(i).valueIN = 100000000 + i
td(i).pHandle = ThreadCreate(pThread, @td(i))
Next I
For i As Integer = 1 To threadNB
ThreadWait(td(i).pHandle)
Next I
t = Timer - t

For i As Integer = 1 To threadNB
Print " SumUpTo(" & td(i).valueIN & ") = " & td(i).valueOUT, _
"(right result : " & (100000000# + i) * (100000000# + i + 1) / 2 & ")"
Next I
Print " total time : " & t & " s"
Print
End Sub

Print
For i As Integer = 1 To 4
Print "Each thread (in parallel) accumulating result directly in shared memory:"
Mythreads(@Thread.SumUpTo_1, I)
Print "Each thread (in parallel) accumulating result internally in its local memory:"
Mythreads(@Thread.SumUpTo_2, I)
Print "-----------------------------------------------------------------------------"
Print
Next i

Sleep

Output (for example):

Each thread (in parallel) accumulating result directly in shared memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
	  total time : 1.668927300015184 s

Each thread (in parallel) accumulating result internally in its local memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
	  total time : 1.004467599958389 s

-----------------------------------------------------------------------------

Each thread (in parallel) accumulating result directly in shared memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
	  total time : 4.314032700025791 s

Each thread (in parallel) accumulating result internally in its local memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
	  total time : 1.032165899962706 s

-----------------------------------------------------------------------------

Each thread (in parallel) accumulating result directly in shared memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
   SumUpTo(100000003) = 5000000350000006  (right result : 5000000350000006)
	  total time : 6.727616399944395 s

Each thread (in parallel) accumulating result internally in its local memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
   SumUpTo(100000003) = 5000000350000006  (right result : 5000000350000006)
	  total time : 1.128656100041894 s

-----------------------------------------------------------------------------

Each thread (in parallel) accumulating result directly in shared memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
   SumUpTo(100000003) = 5000000350000006  (right result : 5000000350000006)
   SumUpTo(100000004) = 5000000450000010  (right result : 5000000450000010)
	  total time : 6.829728199980309 s

Each thread (in parallel) accumulating result internally in its local memory:
   SumUpTo(100000001) = 5000000150000001  (right result : 5000000150000001)
   SumUpTo(100000002) = 5000000250000003  (right result : 5000000250000003)
   SumUpTo(100000003) = 5000000350000006  (right result : 5000000350000006)
   SumUpTo(100000004) = 5000000450000010  (right result : 5000000450000010)
	  total time : 1.164915200012842 s

-----------------------------------------------------------------------------

One can check that the multi-threading performance is strongly penalized by many shared memory accesses in write mode:

For the case where the thread accumulates the result in shared memory, there is no longer any gain from multi-threading (and even a little loss), whereas for the case where the thread accumulates the result in internal memory, the gain is almost at the theoretical maximum value.

On the other hand, we observe a smaller degradation in multi-threading performance when accessing shared memory in read-only mode.

For all pseudo-objects like var-len strings and var-len arrays, only the descriptors can be put into local memory but not the data itself which is always on the heap (only fixed-length data can be put into local memory).
The heap is shared memory and this incurs penality on multi-threading performance as described in the above FAQ.
(see FAQ 16. Why is multi-threading performance penalized by many shared memory accesses (even more in writing mode)?)

For var-len arrays, local fixed-len arrays can be used instead, since this array data is always placed in local scope memory.
Not only do you have to define a fixed maximum size to allocate for each array, but for each of them you have to associate an index variable (per dimension if necessary) that points to the last useful element ('Redim' is replaced by updating this index variable).

For var-len strings, local fix-len [z]strings can be used instead, since these [z]string data are always placed in local scope memory.
All built-in string functions except 'Len()' and 'Asc()' and all string operators should also not be used on [z]strings since they work internally with var-len strings. Instead, use user code that only works on [z]string indexes.
Note:

But fix-len strings ('Dim As String * N') are less convenient to use than fix-len zstrings ('Dim As Zstring * N'), because the former cannot be passed by reference to a procedure (but only by copy), unlike the latter ('Byref As Zstring').

Additionally, all dynamic memory allocation/reallocation/deallocation requests (to be thread-safe) are serialized internally using mutex locking and unlocking.

The following example compares the multithreaded performance of two types of code:

- code with var-len strings using its built-in functions and operators like '= (assign)', 'Instr()', 'Mid()' and 'Ucase()',
- code with fix-len zstrings with user code equivalent to the previous built-in functions and operators, but operating only on zstring indexes.

("Asc()" and "Len()" are the only ones used because they have no impact on performance)

Type Thread
Dim As UInteger value
Dim As Any Ptr pHandle
Declare Static Sub thread1(ByVal pt As Thread Ptr)
Declare Static Sub thread2(ByVal pt As Thread Ptr)
End Type

Sub Thread.thread1(ByVal pt As Thread Ptr)
Dim As Integer result
For n As Integer = 1 To pt->value
Dim As String s1
Dim As String s2
Dim As String s3
s1 = "FreeBASIC rev 1.20"
result = InStr(s1, "rev")
s2 = Mid(s1, result)
s3 = UCase(s2)
Next n
End Sub

Sub Thread.thread2(ByVal pt As Thread Ptr)
Dim As Integer result
For n As Integer = 1 To pt->value
Dim As ZString * 256 z1
Dim As ZString * 256 z2
Dim As ZString * 256 z3
' instead of: z1 = "FreeBASIC rev 1.20"
For i As Integer = 0 To Len("FreeBASIC rev 1.20")
z1[i] = ("FreeBASIC rev 1.20")[i]
Next i
' instead of: result = Instr(z1, "rev")
result = 0
For i As Integer = 0 To Len(z1) - Len("rev")
For j As Integer = 0 To Len("rev") - 1
If z1[i + j] <> ("rev")[j] Then Continue For, For
Next j
result = i + 1
Exit For
Next i
' instead of: z2 = Mid(z1, result)
For i As Integer = result - 1 To Len(z1)
z2[i - result + 1] = z1[i]
Next i
' instead of: z3 = Ucase(z2)
For i As Integer = 0 To Len(z2)
z3[i] = z2[i]
If z3[i] >= Asc("a") Andalso z3[i] <= Asc("z") Then z3[i] -= 32
Next i
Next n
End Sub

Sub MyThreads(ByVal pThread As Any Ptr, ByVal threadNB As UInteger = 1)
Dim As Thread td(1 To threadNB)
Dim As Double t

t = Timer
For i As Integer = 1 To threadNB
td(i).value = 100000
td(i).pHandle = ThreadCreate(pThread, @td(i))
Next I
For i As Integer = 1 To threadNB
ThreadWait(td(i).pHandle)
Next I
t = Timer - t

Print " total time for " & threadNB & " threads in parallel: " & t & " s"
Print
End Sub

Print
For i As Integer = 1 To 8
Print "Each thread using var-len strings, with its built-in functions and operators:"
Mythreads(@Thread.thread1, I)
Print "Each thread using fix-len zstrings, with user code working on zstring indexes:"
Mythreads(@Thread.thread2, I)
Print "------------------------------------------------------------------------------"
Print
Next i

Sleep

Output (for example):

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 1 threads in parallel: 0.08449090004432946 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 1 threads in parallel: 0.02201449999120086 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 2 threads in parallel: 0.1947050000308082 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 2 threads in parallel: 0.02090729994233698 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 3 threads in parallel: 0.3338784999214113 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 3 threads in parallel: 0.0279372000368312 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 4 threads in parallel: 0.4927077000029385 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 4 threads in parallel: 0.02361949998885393 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 5 threads in parallel: 0.7089884000597522 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 5 threads in parallel: 0.02638950000982732 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 6 threads in parallel: 0.9172402999829501 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 6 threads in parallel: 0.0310587000567466 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 7 threads in parallel: 1.159198799985461 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 7 threads in parallel: 0.02898070006631315 s

------------------------------------------------------------------------------

Each thread using var-len strings, with its built-in functions and operators:
	  total time for 8 threads in parallel: 1.403980100061744 s

Each thread using fix-len zstrings, with user code working on zstring indexes:
	  total time for 8 threads in parallel: 0.03312029992230237 s

------------------------------------------------------------------------------

One can check that the multi-threading performance is strongly penalized by many var-len string manipulation:

For the case where the thread uses var-len strings and its built-in functions and operators, there is no longer any gain from multi-threading (and even losses), whereas for the case where the thread uses fix-len zstrings and user code working on zstring indexes only (except 'Asc()' and 'Len()' usage), the gain is almost at the theoretical maximum value.

Critical Sections FAQ - FreeBASIC Wiki Manual | FBWiki

'Critical sections' related questions

'Critical sections' related questions