in reply to Perl Threads Question.

My question is (Once i remove the sleeps) does the below method make 10 calls concurrently at the same time and then return the data for all 10 children. Or will this method simply make 10 calls one at a time defeating the purpose handily.

I'm not sure why you've posted the fork code if your question is about threads? But....

Your threaded code can be reduced (for brevity) to:

#! perl -slw use strict; use threads; my @threads = map async( sub{ my $num = shift; print "started thread $num"; sleep $num; print "done with thread $num"; return $num; }, $_ ), 1 .. 3; print "Done with ", $_->join for @threads; __END__ C:\test>junk52 started thread 1 started thread 2 started thread 3 done with thread 1 Done with 1 done with thread 2 Done with 2 done with thread 3 Done with 3

Now my guess is that your real question is: why do all three threads finish in the same order as I started them.

The answer is: if you start thread 1 and ask it to sleep for 1 second; and start thread 2 and ask it to sleep for 2 seconds; and thread 3 for 3 seconds; and so on. Then the threads will inevitably finish in the same order you started them in.

If you want to reassure yourself that threads actually do run concurrently, you should ask them to sleep for random numbers of seconds:

#! perl -slw use strict; use threads; my @threads = map async( sub{ my( $num, $sleep ) = @_; print "started thread $num to sleep for $sleep seconds"; sleep $sleep; print "done with thread $num"; return $num; }, $_, 2+int( rand 3 ) ), 1 .. 3; print "Done with ", $_->join for @threads; __END__ C:\test>junk52 started thread 1 to sleep for 4 seconds started thread 2 to sleep for 3 seconds started thread 3 to sleep for 2 seconds done with thread 3 done with thread 2 done with thread 1 Done with 1 Done with 2 Done with 3

Now you can see (from the ordering of the "done with ..." messages), that the order they complete in depends upon how long they run. And if you time them, you'll see that the overall time is the largest $sleep value, not the sum of all the $sleep values, as it would be if they ran serially.

Finally, you're probably going to ask why, if thread 3 finished first, why did you get it results last?

The answer is, because that's what was asked for.

Just as in your original code when you pushed the thread handles in the order 1 .. 10 and then asked for the results (joined them) by iterating over that array in the same order.

When the above code does:

print "Done with ", $_->join for @threads;

It is asking for the results of threads in the same order the threads were started, regardless of the order in which they produced those results.


Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
RIP an inspiration; A true Folk's Guy

Replies are listed 'Best First'.
Re^2: Perl Threads Question.
by Monkomatic (Sexton) on Oct 05, 2010 at 02:21 UTC

    Thank you for your answer.

    "I'm not sure why you've posted the fork code if your question is about threads? But...."

    I had thought every fork was a thread.... And a call to fork() started a child process and returned the PID of that process.

    I had heard fork() leakes like a sieve even in 12. so a threads version would be very helpful. However I am not proficient enough to follow your code completely.

    #! perl -slw use strict; use threads; my @threads = map async( sub{ my( $num, $sleep ) = @_; print "started HTTP CALL"; print "done with HTTP CALL return results";return $num; }, print "Done with 10 HTTP CALLS RETURN RESULTS", $_->join for @threads; # I am assuming @threads now holds the 10 results for each child proce +ss. Correct?
    download
      I had thought every fork was a thread.... And a call to fork() started a child process and returned the PID of that process.

      fork starts a process. Not a thread.

      A process consists of at least one thread and can contain multiple threads.

      Each process has its own address space. The threads of a single process share that same address space.

      I had heard fork() leakes like a sieve even in 12.

      Where did you hear that?

      If it is true, and I'm not aware that it is, it will (almost certainly) be platform dependant. Which OS are you using?

      However I am not proficient enough to follow your code completely.... I am assuming @threads now holds the 10 results for each child process. Correct?

      No. @threads contains handles to the 10 threads (objects). The values returned by the threads--if any--are not accessible to you until you call the method join() on those threads handles (objects). The values returned from the threads are stored internally until you fetch them using the join() method.

      BTW: Embedding part of your question within the <code></code> tags is not helpful. They can easily be missed. Please use code tags for code; not for (parts of) the question.


      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.

        I had posted a lengthy Thank you for your response but it never posted for some reason. So im posting it again.

        Thank you much that cleared it all up for me.

        especially the stored internally untill join() part.