MCE::Hobo - A threads-like parallelization module
This document describes MCE::Hobo version 1.699_012
use MCE::Hobo; MCE::Hobo->create( sub { print "Hello from hobo\n" } )->join(); sub parallel { my ($arg1) = @_; print "Hello again, $arg1\n"; } MCE::Hobo->create( \¶llel, $_ ) for 1 .. 3; my @hobos = MCE::Hobo->list(); my @running = MCE::Hobo->list_running(); my @joinable = MCE::Hobo->list_joinable(); my @count = MCE::Hobo->pending(); $_->join() for @hobos; my $hobo = mce_async { foreach (@files) { ... } }; $hobo->join(); if (my $err = $hobo->error()) { warn("Hobo error: $err\n"); } # Get a hobo's object $hobo = MCE::Hobo->self(); # Get a hobo's ID $tid = MCE::Hobo->tid(); # "$$.tid" $tid = $hobo->tid(); $pid = MCE::Hobo->pid(); # $$ $pid = $hobo->pid(); # Test hobo objects if ($hobo1 == $hobo2) { ... } # Give other hobos a chance to run MCE::Hobo->yield(); MCE::Hobo->yield(0.05); # Return context, wantarray aware my ($value1, $value2) = $hobo->join(); my $value = $hobo->join(); # Check hobo's state if ($hobo->is_running()) { sleep 1; } if ($hobo->is_joinable()) { $hobo->join(); } # Send a signal to a hobo $hobo->kill('SIGUSR1'); # Exit a hobo MCE::Hobo->exit();
A hobo is a migratory worker inside the machine that carries the asynchronous gene. Hobos are equipped with threads-like capability for running code asynchronously. Unlike threads, each hobo is a unique process to the underlying OS. The IPC is managed by MCE::Shared, which runs on all major platforms including Cygwin.
threads
MCE::Shared
MCE::Hobo may be used as a standalone or together with MCE including running alongside threads.
MCE::Hobo
MCE
The following is a parallel demonstration.
use strict; use warnings; use MCE::Hobo; use MCE::Shared Sereal => 1; # Serialization via Sereal if available. use MCE::Shared::Ordhash; # Ordered hash for non-shared use below. # usage: head -20 file.txt | perl script.pl # my $aa = MCE::Shared->array(); # OO with on-demand deref @{} # my $ha = MCE::Shared->hash(); # OO with on-demand deref %{} # my $oh = MCE::Shared->ordhash(); # OO with on-demand deref %{} # my $va = MCE::Shared->scalar(); # OO interface only: ->set, ->get my $ifh = MCE::Shared->handle( "<", \*STDIN ); # shared my $ofh = MCE::Shared->handle( ">", \*STDOUT ); my $ary = MCE::Shared->array(); sub parallel_task { my ( $id ) = @_; while ( <$ifh> ) { printf {$ofh} "[ %4d ] %s", $., $_; # $ary->[ $. - 1 ] = "[ ID $id ] read line $.\n" ); # dereference $ary->set( $. - 1, "[ ID $id ] read line $.\n" ); # OO faster } } my $hobo1 = MCE::Hobo->new( "parallel_task", 1 ); my $hobo2 = MCE::Hobo->new( \¶llel_task, 2 ); my $hobo3 = MCE::Hobo->new( sub { parallel_task(3) } ); $_->join for MCE::Hobo->list(); # search array -- single IPC my @vals = $ary->vals( "val =~ / ID 2 /" ); print {*STDERR} join( "", @vals );
The following list options which may be overridden when loading the module.
use Sereal qw( encode_sereal decode_sereal ); use CBOR::XS qw( encode_cbor decode_cbor ); use JSON::XS qw( encode_json decode_json ); use MCE::Hobo freeze => \&encode_sereal, ## \&Storable::freeze thaw => \&decode_sereal ## \&Storable::thaw ;
There is a simpler way to enable Sereal. The following will attempt to use Sereal if available, otherwise defaults to Storable for serialization.
use MCE::Hobo Sereal => 1; use MCE::Shared Sereal => 1; # <-- supports Sereal only at this time
This will create a new hobo that will begin execution with function as the entry point, and optionally ARGS for list of parameters. It will return the corresponding MCE::Hobo object, or undef if hobo creation failed.
Options may be specified via a hash structure. At this time, posix_exit is the only option supported which calls POSIX::_exit(0) when finished. The default is CORE::exit(0). Set posix_exit to avoid all END and destructor processing.
posix_exit
POSIX::_exit(0)
CORE::exit(0)
FUNCTION may either be the name of a function, an anonymous subroutine, or a code ref.
my $hobo = MCE::Hobo->create( "func_name", ... ); # or my $hobo = MCE::Hobo->create( sub { ... }, ... ); # or my $hobo = MCE::Hobo->create( \&func, ... );
The new() method is an alias for create().
new()
create()
mce_async runs the block asynchronously similarly to MCE::Hobo-create()>. It returns the hobo object, or undef if hobo creation failed.
mce_async
MCE::Hobo-
my $hobo = mce_async { foreach (@files) { ... } }; $hobo->join(); if (my $err = $hobo->error()) { warn("Hobo error: $err\n"); }
This will wait for the corresponding hobo to complete its execution. In non-voided context, join() will return the value(s) of the entry point function.
join()
The context (void, scalar or list) for the return value(s) for join is determined at the time of joining and mostly wantarray aware.
join
wantarray
my $hobo1 = MCE::Hobo->create( sub { my @results = qw(foo bar baz); return (@results); }); my @res1 = $hobo1->join; # ( foo, bar, baz ) my $res1 = $hobo1->join; # baz my $hobo2 = MCE::Hobo->create( sub { return 'foo'; }); my @res2 = $hobo2->join; # ( foo ) my $res2 = $hobo2->join; # foo
Tests if two hobo objects are the same hobo or not. Hobo comparison is based on process IDs. This is overloaded to the more natural forms.
if ($hobo1 == $hobo2) { print("Hobos are the same\n"); } # or if ($hobo1 != $hobo2) { print("Hobos differ\n"); }
Hobos are executed in an eval context. This method will return undef if the hobo terminates normally. Otherwise, it returns the value of $@ associated with the hobo's execution status in its eval context.
eval
undef
$@
This sends 'SIGQUIT' to the hobo object, notifying hobo to exit. It returns the hobo object to allow for method chaining. It is important to join later if not immediately to not leave a zombie or defunct process.
'SIGQUIT'
$hobo->exit()->join(); ... $hobo->join(); # later
A hobo can be exited at any time by calling MCE::Hobo-exit()>. This behaves the same as exit(status) when called from the main process.
exit(status)
This class method is called automatically by END, but may be called explicitly. Two shared objects to MCE::Shared are destroyed. An error is emitted via croak if there are active hobos not yet joined.
END
MCE::Hobo->create( 'task1', $_ ) for 1 .. 4; $_->join for MCE::Hobo->list(); MCE::Hobo->create( 'task2', $_ ) for 1 .. 4; $_->join for MCE::Hobo->list(); MCE::Hobo->create( 'task3', $_ ) for 1 .. 4; $_->join for MCE::Hobo->list(); MCE::Hobo->finish();
Returns true if a hobo is still running.
Returns true if the hobo has finished running and not yet joined.
Sends the specified signal to the hobo. Returns the hobo object to allow for method chaining. As with exit, it is important to join eventually if not immediately to not leave a zombie or defunct process.
exit
$hobo->kill('SIG...')->join();
The following is a parallel demonstration comparing MCE::Shared against Redis and Redis::Fast on a Fedora 23 VM. Joining is after all workers have been notified to quit.
Redis
Redis::Fast
use Time::HiRes qw(time); use Redis; use Redis::Fast; use MCE::Hobo; use MCE::Shared; my $redis = Redis->new(); my $rfast = Redis::Fast->new(); my $array = MCE::Shared->array(); sub parallel_redis { my ($_redis) = @_; my ($count, $quit, $len) = (0, 0); # instead, use a flag to exit loop $SIG{'QUIT'} = sub { $quit = 1 }; while (1) { $len = $_redis->rpush('list', $count++); last if $quit; } $count; } sub parallel_array { my ($count, $quit, $len) = (0, 0); # do not exit from inside handler $SIG{'QUIT'} = sub { $quit = 1 }; while (1) { $len = $array->push($count++); last if $quit; } $count; } sub benchmark_this { my ($desc, $num_hobos, $timeout, $code, @args) = @_; my ($start, $total) = (time(), 0); MCE::Hobo->new($code, @args) for 1..$num_hobos; sleep $timeout; # joining is not immediate; ok $_->kill('QUIT') for MCE::Hobo->list(); # joining later; ok $total += $_->join() for MCE::Hobo->list(); printf "$desc <> duration: %0.03f secs, count: $total\n", time() - $start; sleep 0.2; } benchmark_this('Redis ', 8, 5.0, \¶llel_redis, $redis); benchmark_this('Redis::Fast', 8, 5.0, \¶llel_redis, $rfast); benchmark_this('MCE::Shared', 8, 5.0, \¶llel_array);
Returns a list of all hobos not yet joined.
Returns a list of all hobos that are still running.
Returns a list of all hobos that have completed running. Thus, ready to be joined without blocking.
Returns a count of all hobos not yet joined.
Class method that allows a hobo to obtain it's own MCE::Hobo object.
Returns the ID of the hobo. TID is composed of process and thread IDs together as a string value.
PID: $$ TID: "$$.tid"
Class methods that allows a hobo to obtain its own ID.
Let this hobo yield CPU time to other hobos. By default, the class method calls sleep(0.0002) on Unix including Cygwin and sleep(0.001) on Windows.
sleep(0.0002)
sleep(0.001)
MCE::Hobo->yield(); MCE::Hobo->yield(0.05);
The inspiration for MCE::Hobo comes from wanting threads-like behavior for processes. Both can run side-by-side including safe-use by MCE workers. Likewise, the documentation resembles threads.
forks
forks::BerkeleyDB
Thread::Tie
MCE, MCE::Core, MCE::Shared
Mario E. Roy, <marioeroy AT gmail DOT com>
To install MCE, copy and paste the appropriate command in to your terminal.
cpanm
cpanm MCE
CPAN shell
perl -MCPAN -e shell install MCE
For more information on module installation, please visit the detailed CPAN module installation guide.