As brother poolpi has indicated, your code is an ugly mish mosh of threads and process forking...

...assuming mean threads then the code can be reduced somewhat:

use strict ; use warnings ;

use threads ;

my @array = qw(a b c d e f g h);
my @children;
my @results;

for (1..5) {
  for (1..2) {
    push @children, threads->create(\&sub_thread, @array) ;
  } ;

  while (@children) {
    my $child = shift(@children) ;
    push @results, $child->join() ;
  } ;
} ;

sub sub_thread {
  my (@temp) = (@_) ;

  my $tid = threads->tid() ;
  print "child $tid: @temp\n" ;

  sleep (2) ;

  return 'whatever' ;
} ;
[download]

where:

• no semaphore is required. The main thread creates two child threads (note that there's no need to create an anonymous subroutine to do that), and then waits for them both to return (using $child->join()).

  This example shows the thread returning a single scalar. It is important to understand that when a thread is started it receives a copy of the parent's state (including all variables) -- and once started, it cannot change the parent's state. The thread can return stuff to the parent as it terminates, but if you want any other communication between thread and parent (or other thread) you need shared variables or queues or semaphores or ...

  Note especially that the context in which the threads->create() is called is implicity the context for the return from the thread -- unless you use the explicit context setting form. In this case the push actually provides a list context... even though it is returning a single value. (This works, but you may wish to be less sloppy !) Beware, however, of calling threads->create() in void context !

• the thread subroutine does not need to fork. Note that the parameters passed to the thread subroutine are the parameters passed to threads->create() less the subroutine reference parameter (the first parameter).

• no sleeping or yielding is required in the parent, the $child->join() will block the parent until the next child returns.

  Note that the way this is done means that the parent will wait for the children to complete (a) indefinitely, and (b) in the order they were created. If the second child finishes first, the parent will still wait for the first to finish.

If you want to have two threads on the go at the same time at all times, then you need a mechanism for each thread to signal to the parent that it has finished. One way of doing that is a queue. For example:

use strict ; use warnings ;

use threads ;
use Thread::Queue ;

my @array = qw(a b c d e f g h);
my @results;

my $finished_queue = new Thread::Queue ;

my $c = 2 ;     # Number active at once
my $n = 10 ;    # Total number to dispatch
my $m = $n ;    # Number to collect

while ($m) {
  while ($n && $c) {
    my $child = threads->create(\&sub_thread, @array) ;
    $n-- ;
    $c-- ;
  } ;

  my $tid = $finished_queue->dequeue() ;
  my $child = threads->object($tid) ;
  push @results, $child->join() ;

  $m-- ;
  $c++ ;
} ;

sub sub_thread {
  my (@temp) = (@_) ;

  my $tid = threads->tid() ;
  print "child $tid: @temp\n" ;

  sleep(rand(2)+2) ;

  $finished_queue->enqueue($tid) ;

  print "child $tid: finished\n" ;
  return "whatever from child $tid" ;
} ;
[download]

noting: (a) that the order of the results is not necessarily the same as the order of dispatch (which requires some extra information related to the $tid); and (b) that even though it doesn't need the $child while it is creating each one, it does need to set scalar context !

Finally, the other way to run two threads is to start two and pass "parameters" and results back and forth using two queues... along the lines of:

use strict ; use warnings ;

use threads ;
use Thread::Queue ;

my @array = qw(a b c d e f g h);
my @results;

my $param_queue  = new Thread::Queue ;
my $result_queue = new Thread::Queue ;

my $c =  2 ;    # concurrent threads
my $n = 10 ;    # amount of work
my @children ;

for (1..$c) {   # Start the threads
  push @children, threads->create(\&sub_thread) ;
} ;  

for (1..$n) {   # Dispatch the work
  $param_queue->enqueue([@array]) ;
} ;

for (1..$n) {   # Collect the results
  push @results, $result_queue->dequeue() ;
} ;

$param_queue->enqueue((undef) x $c) ;  # Signal the threads to stop

foreach (@children) {   # Collect the terminated threads
  $_->join() ;
} ;

sub sub_thread {
  my $task = 0 ;
  my $tid = threads->tid() ;

  print "child $tid: started" ;

  while (my $p = $param_queue->dequeue()) {
    $task++ ;
    my @temp = @$p ;

    print "child $tid/$task: @temp\n" ;

    sleep(rand(2)+2) ;

    $result_queue->enqueue("whatever from child $tid/$task") ;

    print "child $tid/$task: done\n" ;
  } ;

  print "child $tid: terminated\n" ;
} ;
[download]

which I hope is reasonably self explanatory.

---

---

Posts are HTML formatted. Put <p> </p> tags around your paragraphs. Put <code> </code> tags around your code and data!

Titles consisting of a single word are discouraged, and in most cases are disallowed outright.

Read Where should I post X? if you're not absolutely sure you're posting in the right place.

Please read these before you post! —

• How do I compose an effective node title?
• How do I post a question effectively?
• Markup in the Monastery

Posts may use any of the Perl Monks Approved HTML tags:

a, abbr, b, big, blockquote, br, caption, center, col, colgroup, dd, del, details, div, dl, dt, em, font, h1, h2, h3, h4, h5, h6, hr, i, ins, li, ol, p, pre, readmore, small, span, spoiler, strike, strong, sub, summary, sup, table, tbody, td, tfoot, th, thead, tr, tt, u, ul, wbr

You may need to use entities for some characters, as follows. (Exception: Within code tags, you can put the characters literally.)

	For:		Use:
	&		&
	<		&lt;
	>		&gt;
	[		&#91;
	]		&#93;

Link using PerlMonks shortcuts! What shortcuts can I use for linking?

See Writeup Formatting Tips and other pages linked from there for more info.