Rand Stats

Seq::PreFetch

cpan:SAMGWISE

NAME

Seq::PreFetch - Asynchronously pre-fetch the next item of a Seq

SYNOPSIS

use Seq::PreFetch;

sub slow-and-lazy( --> Seq) {
  gather for 1..* {
    # a time expensive option... like sleep
    sleep 0.5;
    .take
  }
}

my $moment = now;
for slow-and-lazy.&pre-fetch {
  .say;
  say "Delta: { now - $moment }";
  $moment = now;
  sleep 1;
}

DESCRIPTION

Seq::PreFetch asynchronously pre-fetches the next item of a Seq before you pull from the Seq. It provides the sub pre-fetch which wraps a Seq with a pre-fetching Seq.

This pattern allows you to consume one value from a Seq and begin concurrently calculating the next value of a Seq ready for the next time you need a value. Time efficiency gains can be made for operations where the time cost of a consuming loop is greater than the cost of starting a thread to pre-fetch.

As demonstrated in the synopsis example, The two operations do not block so the longer sleep in the loop defines the duration after the first iteration.

AUTHOR

Sam Gillespie samgwise@gmail.com

COPYRIGHT AND LICENSE

Copyright 2017 Sam Gillespie

This library is free software; you can redistribute it and/or modify it under the Artistic License 2.0.