Re^9: Why should any one use/learn Perl 6?


Just another Perl shrine
	PerlMonks

Re^9: Why should any one use/learn Perl 6?

by liz (Monsignor)

on Jun 13, 2018 at 22:26 UTC ( [id://1216598]=note: print w/replies, xml )

Need Help??

in reply to Re^8: Why should any one use/learn Perl 6?
in thread Why should any one use/learn Perl 6?

Thank you for your elaboration.

If I understand you correctly, you either have a little boilerplate with tie

tie my $var, 'MCE::Shared', 0;
$var = 42;
[download]

or you have a lot of boilerplate without tie

my $var = MCE::Shared->share( { module => 'My::Scalar' }, 0 );
$var->set(42)
[download]

And underneath you have locking going on in either case. Is that a correct summary?

Please note that I really like what you have done with MCE!

Comment on Re^9: Why should any one use/learn Perl 6? Select or Download Code

Replies are listed 'Best First'.

Re^10: Why should any one use/learn Perl 6?
by marioroy (Prior) on Jun 13, 2018 at 23:29 UTC

Hello liz,

Under the hood (MCE::Shared::Server), the boilerplate is the same between OO and TIE for the most part. For Perl-like behavior, the TIE interface involves additional overhead from Perl itself.

MCE::Mutex is used internally for locking. There are 10 locks, one per data channel. A worker from MCE, MCE::Hobo, or threads is assigned a data-channel automatically in a round-robin fashion. 10 workers max can obtain an exclusive lock simultaneously. The unidirectional command-channel involves no locking.

Other parallel modules on MetaCPAN may benefit from multiple data channels simply by calling MCE::Shared::init() inside the worker.

Regards, Mario

[reply]

In Section Seekers of Perl Wisdom