Revision history for Perl module MCE.
1.879 Tue May 24 05:00:00 EST 2022
* Replace http with https in documentation and meta files.
* Call PDL::set_autopthread_targ(1); disables PDL auto-threading.
1.878 Sun Feb 20 06:45:00 EST 2022
* Fix for the fast channel implementations.
Thank you, twata for the test report.
1.877 Sun Feb 20 02:30:00 EST 2022
* Improved suppressing the PDL CLONE warning. Piddles should not be
naively copied into new threads.
* Added fast channel implementations optimized for non-Unicode strings.
The main difference is that these lack freeze-thaw serialization.
MCE::Channel::MutexFast, MCE::Channel::SimpleFast, and
MCE::Channel::ThreadsFast.
1.876 Thu Dec 02 18:00:00 EST 2021
* Allow percentage above 100% for max_workers in MCE.
* MCE::Child update.
Improved _ordhash.
Renamed JOINED to REAPED in code for better clarity.
Specify a percentage for max_workers.
Added t/05_mce_child_max_workers.t
1.875 Tue Nov 16 04:00:00 EST 2021
* Specify a percentage for max_workers.
Thank you, kcott@PerlMonks (Ken) for the idea.
https://www.perlmonks.org/?node_id=11134439
* Added t/03_max_workers.t
1.874 Tue Aug 18 16:00:00 EST 2020
* Improved MCE->yield when used together with MCE::Relay.
1.873 Sat Aug 01 16:00:00 EST 2020
* Removed unused variable in MCE::Mutex::Channel.
* Fixed typo in prior change log.
Replaced "completed" with "feature complete".
1.872 Sun Jun 14 22:30:00 EST 2020
* Added open to required dependencies.
* Set default encodings on standard filehandles in tests using UTF-8.
* Bumped minimal Perl version to 5.8.1.
* Bumped MCE version to 1.872 to align with MCE::Shared.
* The MCE project is feature complete.
1.868 Sun May 10 22:00:00 EST 2020
* Completed threads-like detach capability in MCE::Child.
* Resolved MCE::Channel failing when calling dequeue multiple times
on an ended channel.
* MCE->say, MCE->print, and MCE->printf now return 1.
1.867 Sun May 03 18:00:00 EST 2020
* Bug fix for UTF-8 issues during inter-process communication.
This update required undoing optimizations specific to scalar args.
Essentially, IPC involves serialization for everything going forward.
Install Sereal::Encoder and Sereal::Decoder for better performance
in Perl 5.8.8+.
* MCE options flush_stdout, flush_stderr, and flush_file now default to
enabled for the MCE->print, MCE->printf, and MCE->say output routines.
* Improved MCE::Child with threads-like detach capability. See POD.
* Improved IPC in MCE::Queue with permanent fast-like dequeue including
dequeue_nb. Going forward, the fast and barrier options are silently
ignored if specified (i.e. no-op).
* Improved IPC performance on Linux.
1.866 Sat Feb 08 21:00:00 EST 2020
* Bug fix for restart_worker, race condition introduced in 1.863.
Thank you, Oliver Gorwits for reporting the issue.
1.865 Wed Dec 25 18:00:00 EST 2019
* Bug fix for two-way IPC stalling on Windows in MCE::Channel::Threads.
See https://www.perlmonks.org/?node_id=11110612
* Remove the check for MSWin32 in MCE::Channel::Mutex. MCE::Channel since
the 1st release silently defaults to MCE::Channel::Threads on Windows.
* Small tweak to MCE::Signal.
1.864 Wed Dec 04 13:00:00 EST 2019
* Bug fix in MCE::Signal - Shared manager not exiting, introduced in 1.863.
* Use monotonic clock if available in MCE->yield and MCE::Child->yield.
See https://www.perlmonks.org/?node_id=11109673
1.863 Sun Nov 26 20:00:00 EST 2019
* On Cygwin, silently use Mutex in MCE::Channel when Threads is specified
for better performance.
* New defer capability in MCE::Signal. This applies to MCE::Shared 1.863.
See POD section labled "DEFER SIGNAL" in MCE::Signal.
* Reverted $child->exit back to sending the SIGQUIT signal in MCE::Child
now that MCE::Shared::Server 1.863 defers signal during IPC.
* Improved reliability for spawning MCE and MCE::Child inside threads
including nested parallelization, made possible using a global lock
$MCE::_GMUTEX.
* Updated signal handling in mce-examples/framebuffer on GitHub.
1.862 Wed Sep 18 22:00:00 EST 2019
* Hotfix for 1.861.
1.861 Wed Sep 18 08:30:00 EST 2019
* Hotfix for 1.849-1.860. The edge cases regarding signal handling have
been finally resolved for MCE::Child. Thank you, Richard Kelsch for the
use case involving ctrl-c. See mce-examples/framebuffer on GitHub.
1.860 Sun Sep 15 20:00:00 EST 2019
* Signal-handling update release.
* Localized input and output record separators in MCE::Channel.
* IPC safety in MCE::Child during SIGINT and SIGTERM.
* Method $child->exit in MCE::Child now sends the SIGINT signal
for extra reliability with MCE::Shared (previously SIGQUIT).
1.850 Mon Sep 09 12:30:00 EST 2019
* Bumped version to coincide with the stable MCE::Shared 1.850 release.
1.849 Sun Sep 08 23:30:00 EST 2019
* Fixed edge case in MCE::Child when reaping inside a signal handler.
* Added list_pids class method to MCE::Child.
1.848 Tue Sep 03 23:30:00 EST 2019
* Improved IO::All::{ File, Pipe, STDIO } output via MCE->print($io, ...),
printf, and say. This resolves a bug introduced in 1.845 when using
App::Cmd::Tester to capture output.
1.847 Mon Sep 02 23:30:00 EST 2019
* Obsolete RedHat MCE-1.840-Sereal-deps.patch file. This patch file
is no longer needed and finally resolved with this release.
* PDL random numbers now unique between threads. Thank you, PerlMonks (vr).
See https://www.perlmonks.org/?node_id=1214439.
* Replaced "PF_UNIX" with "AF_UNIX" in MCE::Util.
1.846 Mon Aug 26 21:30:00 EST 2019
* Fixed code tags in documentation. Thank you, Mohammad S Anwar.
1.845 Sun Aug 25 22:00:00 EST 2019
* Croak if is_joinable, is_running, or join is called by a non-managed
process in MCE::Child. Added LIMITATION section to the documentation.
* Improved is_joinable, is_running, list_joinable, and list_running
in MCE::Child. Thank you, Philippe Baumgart.
* Added example (Consumer requests item) to MCE::Channel documentation.
* Support the task_end option regardless if user_tasks is specified.
* Support IO::All::{ File, Pipe, STDIO } for input data including
output via MCE->print($io, ...), printf, and say.
* Support gather => MCE::Candy::out_iter_fh($io) using MCE::Candy.
1.844 Wed Aug 14 21:30:00 EST 2019
* Resolved MCE stalling when specifying max_retries with init_relay.
Ditto for loop_timeout with init_relay on UNIX platforms.
Thank you, Chris Denley.
* Enhanced loop_timeout to handle workers dieing uncontrollably from
any user_tasks (i.e. task_id >= 0). Previously, only task_id == 0.
* Improved IPC on the Windows platform for edge case when a worker is
awaiting input while the manager process is restarting a worker.
* MCE, MCE::Child workers exit immediately upon receiving a SIGSEGV signal.
This safeguards IPC from stalling inside the manager process.
* Enhanced the _wait_one private function in MCE::Child.
* Removed Prima from the list for auto-enabling the posix_exit option.
Prima (since 1.52) is parallel safe during global cleanup.
* Reached 100% Pod coverage.
1.843 Tue Jul 23 22:30:00 EST 2019
* Updated results in MCE::Child (Parallel::ForkManager-like demonstration).
* Completed missing interrupt signal-safety for the non-blocking methods
in MCE::Channel::Mutex and MCE::Channel::Threads.
1.842 Sun Jul 21 19:00:00 EST 2019
* Fixed race condition abnormalities in MCE::Child.
* Added Parallel::ForkManager-like demonstration to MCE::Child.
1.841 Sun Jul 07 23:30:00 EST 2019
* Disabled t/04_channel_threads testing on Unix platforms for Perl less than
5.10.1. Basically, the MCE::Channel::Threads implementation is not supported
on older Perls unless the OS vendor applied upstream patches (i.e. works on
RedHat/CentOS 5.x running Perl 5.8.x).
* Added LIMITATIONS section to MCE::Channel::Threads.
1.840 Sun Jul 07 05:00:00 EST 2019
* Update MCE::Channel POD documentation.
1.839 Sun Jul 07 04:30:00 EST 2019
[NEW FEATURES]
* Added MCE::Mutex::Channel2 providing two locks using a single channel.
The secondary lock is accessible by calling methods with the '2' suffix.
E.g. primary mutex ->lock, ->unlock; secondary mutex ->lock2, ->unlock2
* Added MCE::Channel providing queue-like and two-way communication
supporting threads and processes.
* Added MCE::Child and compatibility with Perl 5.8. MCE::Child is based on
MCE::Hobo, but using MCE::Channel for data retrieval without involving
a shared-manager process.
* Added MCE::Channel examples { channel1.pl and channel2.pl } using
threads and MCE::Child respectively.
https://github.com/marioroy/mce-examples/tree/master/chameneos
[ENHANCEMENTS]
* IPC update; removed unnecessary overhead including private methods
_sysseek and _syswrite from MCE::Util (no longer needed).
* Improved MCE->do, now callable by workers and the manager process.
* Updated MCE::{ Flow, Grep, Loop, Map, Step, and Stream } documentation
on passing an array reference versus a list for deeply input data.
* Updated and re-organized the top-level MCE documentation, particularly
improved clarity for the 'MCE Models' section.
* Removed MANIFEST.SKIP.
1.838 Wed Jan 23 08:30:00 EST 2019
* IPC update, raising reliability across multiple platforms.
* Improved hack for the Windows platform for nested MCE sessions.
* Added _sysread, _sysseek, _syswrite, and _nonblocking to MCE::Util.
* Added barrier option to MCE::Queue allowing one to disable.
1.837 Sat Aug 25 13:00:00 EST 2018
* Seeds the Math::Random::MT::Auto generator automatically when present
for non-threads, similarly to Math::Random and Math::Prime::Util, to
avoid child processes sharing the same seed value as the parent and
each other. The new seed is computed using the current seed.
1.836 Mon Jun 25 08:30:00 EST 2018
* Moved validation code from MCE::Util to MCE::Core::Validation.
* Applied small optimizations.
1.835 Tue Mar 13 15:00:00 EST 2018
* Added gather and relay demonstrations to MCE::Relay.
* Load IO::Handle for extra stability, preventing workers loading uniquely.
* Load Net::HTTP and Net::HTTPS before spawning if present LWP::UserAgent.
See https://www.perlmonks.org/?node_id=1199760
and https://www.perlmonks.org/?node_id=1199891.
1.834 Tue Jan 23 08:00:00 EST 2018
* Improved Queue await and dequeue performance on the Windows platform.
* Added chameneos-redux parallel demonstrations on GitHub:
https://github.com/marioroy/mce-examples/tree/master/chameneos
1.833 Thu Dec 28 16:00:00 EST 2017
* Fixed bug with sequence (#10), broken in 1.832. Thank you, @djerius.
1.832 Tue Nov 21 15:30:00 EST 2017
* Added LWP::UserAgent to list for enabling posix_exit.
* Improved number-sequence generation for big integers.
* Improved wantarray support in MCE::Mutex synchronize.
* Removed limit check on chunk_size option.
1.831 Sun Oct 08 20:30:00 EST 2017
* Added STFL (Terminal UI) to list for enabling posix_exit.
See https://www.perlmonks.org/?node_id=1200923.
* Math::Prime::Util random numbers now unique between MCE workers.
See https://www.perlmonks.org/?node_id=1200960.
1.830 Tue Sep 12 17:00:00 EST 2017
[BUG FIXES]
* Fixed MCE and MCE::Relay stalling when setting the input record separator.
See https://www.perlmonks.org/?node_id=1196701. Thank you, JediWombat.
* Fixed bug with dequeue_nb in MCE::Queue (#8). Thank you, @bokutin.
* Fixed signal handler (#9). Thank you, @chrisdenley.
[ENHANCEMENTS]
* Added Coro and Win32::GUI to list for enabling posix_exit.
* Added support for Haiku to get_ncpu in MCE::Util.
* Allow gathering to a shared array in MCE::Candy.
* Improved CPU count on the AIX platform in MCE::Util.
* Improved signal handling, including nested parallel-sessions.
* Improved running MCE::Hobo inside MCE workers.
* Improved running MCE with PDL.
* Refactored logic for MCE->do, bi-directional callback feature.
* Preserve lexical type for numbers during IPC: MCE->do and MCE::Queue.
* No longer loads threads on the Windows platform in MCE::Signal.
This enables MCE::Hobo 1.827 to spin faster, including lesser
memory consumption. Threads isn't required to run MCE::Hobo.
* Removed extra white-space from POD documentation.
* Validated MCE on SmartOS.
1.829 Wed May 03 03:00:00 EST 2017
* Reduced memory consumption.
1.828 Fri Apr 28 16:00:00 EST 2017
[BUG FIXES]
* Do not enable barrier mode for Queue on the Windows platform.
* Fixed MCE::Mutex::Flock, tmp_file missing script name in path.
[ENHANCEMENTS]
* Added Curses and Prima to list for enabling the posix_exit option.
* Allow a hash as input_data: Core API, MCE::{ Flow, Loop, Step }.
* Improved API documentation on MCE models with more synopsis.
* Enhanced IPC and signal handling. Reduced memory consumption.
* Make tmp_dir on demand in MCE::Signal. Ditto for sess_dir in MCE.
* Load Fcntl, File::Path, Symbol on demand.
1.827 Wed Apr 05 01:30:00 EST 2017
* Do not enable barrier mode for Queue if constructed inside a thread
or by MCE models (e.g. Step, Stream). Ditto for fast => 1 option.
* Updated MCE to not croak when running Perl in taint mode via perl -T.
Failing -T was MCE::Core::Input::{ Generator, Sequence }, MCE::Signal,
and MCE::Util.
* Added Denis Fateyev, Felipe Gasper, and Paul Howarth to Credits.
1.826 Sun Apr 02 23:00:00 EST 2017
* Is now safe running MCE with the Wx GUI toolkit (wxWidgets).
1.825 Sun Apr 02 07:00:00 EST 2017
* Updated MCE::Queue. The following provides a comparison for the
enhancements made regarding IPC during 1.822 through 1.825, in order
to run on machines having "many" cores. I ran with 12, 96, and 192
workers on an 8 core box.
A. MCE queue, dequeue 100k items.
my $Q = MCE::Queue->new();
$Q->enqueue( 1 .. 100000 );
$Q->end(); MCE 1.608 $Q->enqueue((undef) x 12, 96, or 192);
MCE->new(
max_workers => 12, 96, or 192,
user_func => sub {
while ( defined ( my $item = $Q->dequeue ) ) {
;
}
}
)->run();
MCE 1.608: 12 ~ 1.799 secs 96 ~ 8.702 secs 192 ~ 18.083 secs
MCE 1.821: ~ 1.450 secs ~ 5.231 secs ~ 8.102 secs
MCE 1.825: ~ 0.976 secs ~ 1.067 secs ~ 1.509 secs
B. Input file containing 250k lines ( 300 MiB ).
MCE->new(
input_data => "/dev/shm/file_250k.txt",
max_workers => 12, 96, or 192,
chunk_size => 1,
use_slurpio => 1,
user_func => sub { }
)->run();
MCE 1.608: 12 ~ 3.605 secs 96 ~ 8.074 secs 192 ~ 8.465 secs
MCE 1.821: ~ 3.613 secs ~ 8.058 secs ~ 8.607 secs
MCE 1.825: ~ 3.567 secs ~ 2.601 secs ~ 3.199 secs
C. Sequence of numbers from 1 to 200k.
MCE->new(
sequence => [ 1, 200000 ],
max_workers => 12, 96, or 192,
chunk_size => 1,
user_func => sub { }
)->run();
MCE 1.608: 12 ~ 1.236 secs 96 ~ 2.922 secs 192 ~ 3.113 secs
MCE 1.821: ~ 1.227 secs ~ 2.915 secs ~ 3.106 secs
MCE 1.825: ~ 1.250 secs ~ 1.203 secs ~ 1.581 secs
* Results were captured on a fast 8 core system running CentOS Linux 7.
The thing to take from this is that running many workers "no longer"
results in up to 5.6x penalty regarding IPC.
1.824 Sat Apr 01 01:00:00 EST 2017
* Completed validation for running MCE on a box having 100+ cores.
* Tuned the number of data-channels for IPC. Set upper limit in
MCE::Core::Input::{ Handle and Sequence } to not impact the
OS kernel. The result is better performance, yet graceful.
1.823 Fri Mar 31 19:30:00 EST 2017
* Calibrated the number of data-channels for IPC.
1.822 Fri Mar 31 11:00:00 EST 2017
* Check for EINTR during sysread and syswrite.
* Improved reliability when running nested MCE sessions.
* Updated MCE::Mutex with Channel and Fcntl implementations.
* Completed validation for using MCE with 200+ cores.
1.821 Sun Mar 19 04:00:00 EST 2017
* Improved reliability when running MCE with threads.
* Added parallel Net::Pcap and Ping demonstrations on GitHub:
https://github.com/marioroy/mce-examples/tree/master/network
* Optimized 'dequeue' method in MCE::Queue.
* Optimized 'synchronize' method in MCE::Mutex.
1.820 Thu Mar 09 02:00:00 EST 2017
* Improved reliability when running MCE inside an eval block.
1.819 Fri Mar 03 23:00:00 EST 2017
* Fixed issue with localizing AUTOFLUSH variable before autoflush handles.
Thank you, Charles Hendry for raising the issue.
1.818 Wed Mar 01 22:00:00 EST 2017
* Updated bin/mce_grep for determining chunk level. Ditto for chunk size.
Fixed an issue for not seeing STDERR output with '--chunk-level=file'.
Added support for zgrep, zegrep, and zfgrep. Thank you, Jeff Rouse.
* Changed Sereal to Sereal::Decoder and Sereal::Encoder in recommends
section inside Makefile and META files.
* Refactored MCE::Queue. Merged local and manager code base into one.
Removed t/04_norm_que_local.t and t/04_prio_que_local.t.
* Added 'end' method to MCE::Queue.
Updated documentation on dequeue and pending.
1.817 Sat Feb 25 02:00:00 EST 2017
* Improved bin/mce_grep. When -r is specified and zero paths are given,
start recursively in the current directory versus await data from STDIN.
Set chunk-level accordingly to list mode.
1.816 Fri Feb 24 19:00:00 EST 2017
* Revised the description on max_retries in MCE::Core.pod.
1.815 Fri Feb 24 01:00:00 EST 2017
* Fixed divide-by-zero error in MCE->yield.
* Refactored code for the interval option by moving the code to the manager
process. This allows the manager process to accomodate the next available
worker and ready to run. Previoulsy, a worker taking a long time resulted
in empty time slots. Thank you, Philippe Baumgart for your patience.
* Revised the description on posix_exit in MCE::Core.pod.
1.814 Mon Feb 20 05:30:00 EST 2017
* Enhanced the progress option for use with MCE->process.
Updated demonstrations in MCE::Core.pod.
1.813 Thu Feb 16 02:30:00 EST 2017
* Last minute request by Philippe Baumgart (reminder and long overdue).
Added progress option, a code block for receiving info on progress made.
See MCE::Core.pod for demonstrations accommodating all input data types.
1.812 Tue Feb 14 17:00:00 EST 2017
* Bumped minimum requirement for Sereal to 3.015 when available.
Added check ensuring matching version for Encoder and Decoder.
1.811 Mon Feb 13 23:30:00 EST 2017
* Fixed bug in MCE::Queue (dequeue_nb) when queue has zero items.
* Applied small optimization in MCE::Core::Input::Sequence and Generator.
* Added cross-platform template to MCE::Examples for making an executable.
* Removed signal handling for XCPU and XFSZ from MCE::Signal.
* Imply posix_exit => 1 if Gearman::XS or Gearman::Util is present during
MCE construction.
* Added MCE + Gearman demonstrations (xs and non-xs) on GitHub:
https://github.com/marioroy/mce-examples/tree/master/gearman_xs
https://github.com/marioroy/mce-examples/tree/master/gearman
* Changed kilobytes and megabytes to kibiBytes (KiB) and mebiBytes (MiB)
respectively inside the documentation.
1.810 Fri Dec 09 23:30:00 EST 2016
* Updated check for IO handle allowed. This allows $gather_fh = *STDOUT{IO},
construction in Perl <= 5.10.1. Thank you, Qiang Wang.
1.809 Wed Nov 23 16:00:00 EST 2016
* Bug fixes for running MCE inside threads.
* Random numbers are unique between workers.
1.808 Sat Nov 05 02:00:00 EST 2016
* Workers persist unless shutdown explicity while running alongside
the Mojolicious framework.
1.807 Tue Nov 01 16:00:00 EST 2016
* Enhanced relay capabilities. Added Mandelbrot example to MCE::Example.
Added extra demonstrations to MCE::Relay. Also, added test script.
* Tweaked manager-loop delay for special cases -- applies to MSWin32 only.
1.806 Tue Oct 11 21:30:00 EST 2016
* Fixed two typos. Thank you, Florian Schlichting.
* Support input_data with nested arrays in MCE models.
1.805 Thu Sep 01 16:00:00 EST 2016
* Fixed bug in MCE::Queue (#4). Thank you, Mary Ehlers.
* Improved support for running MCE with Tk. Added Tk demonstrations to
MCE::Examples. Thank you, Götz Meyer.
1.804 Thu Jul 28 23:00:00 EST 2016
* Removed the sleep statement in MCE->restart_worker.
* Added FCGI::ProcManager demonstrations to MCE::Examples.
* Automatically set posix_exit to 1 whenever (F)CGI.pm is present.
* Thank you Kai Wasserbäch (TheRealCuran) for the cool test case.
https://github.com/marioroy/mce-perl/issues/1
1.803 Sun Jul 10 23:30:00 EST 2016
* Re-enabled Sereal 3.008+ for Perl < v5.12.0, if available.
* Optimized dequeue methods in MCE::Queue.
1.802 Mon Jul 04 03:30:00 EST 2016
* Default to Storable for serialization in Perl less than v5.12.0.
Sereal 3.008+, if available, is loaded automatically in Perl v5.12+.
1.801 Sun Jul 03 00:30:00 EST 2016
[BUG FIXES]
* Fixed race condition in Queue->await.
[ENHANCEMENTS]
* MCE 1.801 is stable on all supported platforms.
* Completed work supporting cyclical include of MCE Core / models.
* Updated MCE to support Perl included with Git Bash.
* Renamed temp dir from 'mce' to 'Perl-MCE' under user's %TEMP%
location on Windows. E.g. Native Perl, Cygwin, Git Bash.
1.800 Sat Jun 18 16:30:00 EST 2016
[BUG FIXES]
* Fixed dequeue (count) in MCE::Queue for standalone mode.
[ENHANCEMENTS]
* On Windows, improved stablity and feature parity with UNIX.
* Use Sereal 3.008+ automatically if available on the box.
[NEW FEATURES]
* Added support for cyclical include of MCE Core, MCE models, and
MCE Queue by scoping the configuration to the local package.
This resolves (RT#107384), bug reported by Kai Wasserbäch.
1.708 Sat May 28 14:00:00 EST 2016
[BUG FIXES]
* Improved import routine in MCE models and MCE::Subs. This resolves an
issue where functions are not exported; e.g. mce_flow, mce_flow_s.
[ENHANCEMENTS]
* Added support for IO::TieCombine handles. This enables MCE->print and
MCE->sendto to work reliably with App::Cmd and App::Cmd::Tester.
See Testing and Capturing Output in MCE::Examples.
1.707 Wed May 25 16:00:00 EST 2016
[BUG FIXES]
* Fixed logic when workers exit. Improved reliability on Windows.
[ENHANCEMENTS]
* Applied MCE-1.700-provides.patch from RedHat. Thank you, Paul Howarth.
* Added META.json to the distribution.
1.706 Fri Apr 22 21:30:00 EST 2016
[ENHANCEMENTS]
* Time::HiRes sleep resolution is 15 milliseconds on Windows and Cygwin.
Adjusted timeout values accordingly. Thank you, Daniel Dragan.
* Reinstated the hack for faster IO when use_slurpio => 1 is specified.
Tuned chunk_size => 'auto'.
1.705 Thu Apr 14 10:00:00 EST 2016
* Bumped version for Test::More to 0.88. Thank you, Paul Howarth.
1.704 Thu Apr 14 05:00:00 EST 2016
[BUG FIXES]
* Fixed restart on the Windows platform, bug introduced in 1.700.
* Reached *stable* on all major platforms for MCE 1.7x.
[ENHANCEMENTS]
* Enabled auto-destroy for MCE objects.
* Enabled freeze callbacks for Sereal.
* Switched bug tracking to GitHub.
* Tweaked test scripts.
1.703 Sat Mar 19 23:00:00 EST 2016
* Completed IPC optimizations for 1.7.
1.702 Tue Mar 15 17:00:00 EST 2016
* Bumped version.
1.701 Tue Mar 15 12:00:00 EST 2016
[ENHANCEMENTS]
* Some folks have expressed a wish for running MCE 1.7 with Perl v5.8.
To restore support for Perl v5.8, removed MCE::Shared and MCE::Hobo
from the MCE 1.701 distribution.
* MCE::Shared will be released after MCE 1.700 is deleted from CPAN.
Thank you for your patience during this transition.
1.700 Tue Mar 08 15:30:00 EST 2016
[BUG FIXES]
* Fixed race condition on Windows for non-threaded workers.
* Updated MCE models to not fail when running inside an eval statement.
This addresses (RT#105557) and (RT#105559). Thank you, Benjamin McKeown.
* Added new MCE option loop_timeout to prevent the MCE Manager process
from hanging perpetually. The manager process wrongly assumes a worker
is still running when the worker died in an uncontrollable manner.
This resolves (RT#111780). Thank you, Benjamin McKeown.
[ENHANCEMENTS]
* Perl 5.10.1 or later is required to run MCE 1.7. Perl < 5.10.1 lacks
'overloading.pm'.
* Added code in MCE::Grep's documentation for parsing huge files.
* Added support for running MCE with Perl under MobaXterm on Windows.
* MCE/examples and MCE/images are no longer included with the distribution.
These are maintained separately at https://github.com/marioroy/mce-examples
and https://github.com/marioroy/mce-assets respectively.
* MCE performs channel locking via a pipe or socket depending on platform.
Previously, locking was through file locking using flock. This resolves
the slow locking performance on Cygwin.
* Optimized signal handling including improved support on Windows.
* Reduced overhead during spawning and job submissions on Windows and Cygwin.
This enables IPC to complete up to 20x faster, thus benefiting Monte-Carlo
simulations; e.g. calling ->run(0) or ->process(...) repeatedly.
* The MCE::Flow and MCE::Step models can take an anonymous array for
specifying use_threads uniquely for sub-tasks.
[NEW FEATURES]
* Added MCE::Hobo for running code asynchronously. This provides async/join
functionality for processes similarly to async/join in threads. It includes
->is_joinable, ->is_running, ->join, ->kill, ->list, ->waitall, ->waitone,
and other methods not metioned here.
* Added MCE::Shared for sharing objects/data between threads/processes.
* Added MCE::Shared::{ Array, Handle, Hash, Ordhash, and Scalar }.
* Added MCE::Shared::{ Condvar, Minidb, Queue, and Sequence }.
* Added MCE::Shared::{ Server }.
* Added methods ->await, ->enq, and ->enqp to MCE::Step.
* Added method ->await to MCE::Queue.
* Added option max_retries => N for retrying a failed chunk from a worker
dying while processing input data or sequence of numbers.
* Added option posix_exit => 1 to avoid END and destructor processing.
This is necessary for running with Tk and child processes or with
use_threads => 0.
* Seeds the Math::Random generator automatically when present for non-threads
to avoid child processes sharing the same seed value as the parent and
each other. The new seed is computed using the current seed. Thus, okay
to set the seed at the application level for predictable results.
1.608 Fri Apr 10 03:00:00 EST 2015
* Correction for $prog_name (-e) to (perl) in MCE::Signal.
1.607 Fri Apr 10 01:00:00 EST 2015
[BUG FIXES]
* Updated t/01_load_signal_arg.t to address false-positive from bug fix
in 1.606. This was missed in the 1.606 release. Thank you Nigel Horne.
[ENHANCEMENTS]
* Replaced (-e) with (perl) for the $prog_name value in MCE::Signal;
e.g. from running perl -e 'command'.
1.606 Wed Apr 08 18:00:00 EST 2015
[BUG FIXES]
* Added -d and -w tests to ensure $ENV{TEMP} exists and writeable in
MCE::Signal. Otherwise /tmp is used as usual.
[ENHANCEMENTS]
* Determine running state in MCE->exit. Call stop_and_exit if not
already running to not hang the script.
1.605 Mon Apr 06 00:30:00 EST 2015
[BUG FIXES]
* Improved fix for the die handler in MCE::Signal and MCE::Core::Worker.
* Improved support for threads in MCE::Signal's stop_and_exit function.
1.604 Sat Mar 21 21:00:00 EST 2015
[BUG FIXES]
* All bugs found during testing of the upcoming 1.7 release have been
backported to the 1.6 branch.
[NEW FEATURES]
* Added out_iter_array and out_iter_fh to MCE::Candy. These preserve output
order and cover the two general use cases.
1.603 Tue Mar 17 21:00:00 EST 2015
[ENHANCEMENTS]
* A safer solution by Dmitry Karasik for the die handler in MCE::Signal
and MCE::Core::Worker.
* Moved ->forchunk, ->foreach, and ->forseq sugar methods to MCE::Candy.
Stubs exist in MCE. Thus, no breakage to existing apps.
* Removed the link to MCE::Shared on the main page. I decided to backport
all the fixes into 1.6. The MCE::Shared link was missed and requires
the upcoming 1.7 release.
1.602 Mon Mar 16 20:00:00 EST 2015
[BUG FIXES]
* Updated die handler in MCE::Core::Worker and MCE::Signal (RT#102802)
Bug reported by Dmitry Karasik.
[ENHANCEMENTS]
* MCE child processes call CORE::exit during exiting; previously CORE::kill.
* Improved IPC stability on Windows and Cygwin.
* Multiple calls to mutex->lock by the same process or thread is now safe.
The mutex will remain locked until mutex->unlock is called.
1.601 Sun Mar 15 21:00:00 EST 2015
This release is attributed to the many use cases sent by George Bouras.
[BUG FIXES]
* Updated Makefile.PL allowing installation of bin/mce_grep (RT#102040).
Running (MCE_INSTALL_TOOLS=1 perl Makefile.PL) installs bin/mce_grep.
* Ensure MCE instances spawned by workers have shutdown before leaving.
* An exiting forked non-MCE process will not cause the worker to exit.
* Fixed sockets not closing immediately in Cygwin and Windows.
* Fixed a pod error in MCE::Mutex.
* Fixed a rare condition where socket handles for MCE::Mutex and
MCE::Queue were closing pre-maturely when using threads.
* Fixed an undefined variable inside croak statements in MCE models'
import subroutine.
* Fixed automatic shutdown due to loading threads and specifying
use_threads => 0 with workers persisting before exiting the script.
[ENHANCEMENTS]
* Moved relay methods, introduced in 1.600, from MCE to MCE::Relay.
The MCE init_relay option loads and enables MCE::Relay automatically.
* Captured metrics from Linux (previously OS X) for forchunk.pl,
foreach.pl, and forseq.pl. Updated Examples.pod.
* Default to 'auto' for max_workers in bin/mce_grep, examples/egrep.pl,
forchunk.pl, foreach.pl, forseq.pl, iterator.pl, and wc.pl.
* Moved _create_socket_pair from MCE to MCE::Util as _make_socket_pair.
Added _destroy_sockets. Updated MCE, Mutex, and Queue to use MCE::Util.
* Added a CARP_NOT line to MCE models.
* Added support for ->next and ->last from any sub-tasks in MCE::Step.
* Reverted a small change applied in 1.600 to ->do and ->gather.
* The 'synchronize' method in Mutex is wantarray aware.
* Updated POD header lines for method names; from =item to =head2.
* Workers now set STDERR and STDOUT to flush automatically.
[NEW FEATURES]
* Added ->pid method to MCE.
1.600 Sat Jan 31 20:00:00 EST 2015
[BUG FIXES]
* Die handlers (in MCE::Signal and MCE::Core::Worker) are finally 100%.
Furthermore, eval { die ... } statements behave correctly while running
MCE itself inside an eval block. Thus, MCE on iPerl on top of iPython
is possible via Devel::IPerl.
* Fixed queues stalling from running (MCE::Queue fast => 1) on Linux.
* MCE models now set $MCE::FREEZE, $MCE::THAW, and $MCE::TMP_DIR when
overriding freeze, thaw, and tmp_dir respectively at load time.
* Pressing CTRL-D now ends STDIN the first time.
[ENHANCEMENTS]
* Added seven names to CREDITS; David Farrell, Demian Riccardi,
Hisham Eldai, Joel Berger, Oleksandr Kharchenko, Wei Shen, and
Zakariyya Mughal.
* Refactored ->print, ->printf, and ->say. Optimized ->print some more.
* Shorten $_queue to $_Q in Queue. ID is always sent first during IPC.
* The init method in MCE models can now take an array of options.
* Optimized memory consumption in Handle.pm, Iterator.pm, and Request.pm.
* Optimized memory consumption for ->sendto, ->do, and ->gather.
* Optimized memory consumption for overall IPC in general.
* Refreshed the MCE->new method.
[NEW FEATURES]
* Added a new module; MCE::Mutex providing simple semaphore.
* Added ->relay_recv, ->relay, and ->relay_final methods to the Core API.
Refreshed cat.pl and findnull.pl examples to relay the number of lines
read. Workers output exclusively and orderly to STDOUT in cat.pl.
* Added several examples; biofasta (folder), mutex.pl, and relay.pl.
The FASTA examples process by records "\n>", not by lines.
* MCE applies additional logic when RS begins with a newline character;
e.g. RS => "\n>". It trims away characters after the newline and
prepends them to the next record. This happens automatically when not
slurping. Otherwise, the logic is applied to the first and last records
only. This is illustrated in the Core API documentation.
* Updated the Core API documentation (RS, added relay methods).
1.522 Thu Dec 25 16:00:00 EST 2014
[BUG FIXES]
* Applied fix to MCE->shutdown so that MCE models do not err when
receiving signal to terminate.
* Optimization for MCE->print, MCE->printf, MCE->say and MCE->sendto.
This was calling fileno unnecessarily. MCE->print('STDERR', ...)
is not supported. This works; MCE->print(\*STDERR, ...)
[ENHANCEMENTS]
* Inserts the actual lib-path at the head of @INC in example files.
* Massive documentation updates throughout the entire distribution.
* Renamed barrier_sync.pl to sync.pl; scaling_pings.pl to ping.pl.
* Updated requires in META.yml, Makefile.PL and perl-MCE.spec.
* Updated comment in MCE::Queue test scripts.
* Updated the README file.
[NEW FEATURES]
* Added sampledb examples demonstrating DBI and SQLite with MCE.
* Added step_demo.pl to examples.
1.521 Thu Dec 11 16:00:00 EST 2014
[BUG FIXES]
* Fixed broken MCE::Queue ->insert and ->peek methods. FIFO and LIFO
are fully supported with this release.
* Support running in taint mode.
[ENHANCEMENTS]
* Added support for negative index in MCE::Queue ->insert and ->peek.
Updated the documentation.
* CBOR::XS serialization is mentioned in documentation along with
JSON::XS and Sereal.
* Completed code refactoring for the 1.5 branch.
* Optimized argument parsing in import routines.
* Removed the MCE spawn_delay option from test scripts.
[NEW FEATURES]
* Added 15 new test scripts for testing user_args, MCE::Queue
and MCE models.
* An upper-limit of 8 is set when specifying max_workers => 'auto'.
Several folks have requested this. More info at MCE::Util::get_ncpu
for increasing or decreasing the limit.
1.520 Wed Nov 05 03:00:00 EST 2014
[ENHANCEMENTS]
* MCE::Step and MCE::Stream can take the 'fast' option when including
the module. The 'fast' option was introduced in 1.518.
* Removed the type declaration for self (feedback from Sri).
* Removed -s from files_flow.pl (was left behind).
* Added support for GNU Hurd OS in get_ncpu.
1.519 Mon Oct 27 19:00:00 EST 2014
[ENHANCEMENTS]
* Unset the need for channel locking if only worker riding the channel.
There are 8 data channels in MCE. Basically, a worker will obtain a
lock only when sharing the data channel with another worker.
* Updated files_flow.pl, files_mce.pl and files_thr.pl to allow for
many workers for the first task. Updated the synopsis in MCE::Queue.
Synced example listing in Examples.pod with the examples folder.
* Remove period from summary line.
1.518 Mon Oct 27 10:00:00 EST 2014
[BUG FIXES]
* Corrected MCE::Queue's synopsis due to missing List::MoreUtils line.
Changed the synopsis to use 1 worker for the 'dir' task in the event
one were to copy/patse the code and use threads. The glob() function
is not thread-safe in Perl 5.16.x; fixed in 5.18.2, okay in 5.8 - 14.
* Use portable syntax for setpgrp in MCE::Signal. This closes issue 1
at https://code.google.com/p/many-core-engine-perl.
[NEW FEATURES]
* New 'fast' option for MCE::Queue. The 'fast' option speeds up ->dequeue
ops and not enabled by default. It is beneficial for queues not needing
->clear or ->dequeue_nb and not altering the optional count value
while running; e.g. ->dequeue( [ $count ] ).
* Added three examples: files_flow.pl, files_mce.pl, and files_thr.pl.
* Benchmarked on several OSes and appended results to MCE::Queue synopsis.
1.517 Thu Oct 23 10:00:00 EST 2014
[BUG FIXES]
* Correction applied to MCE::Util::get_ncpu for Tru64 UNIX. This method
will emit a warning (not croak) whenever the OS is unknown.
* Changed ${^CHILD_ERROR_NATIVE} to $? in examples/pipe2.pl. This was
missed in the previous release.
[ENHANCEMENTS]
* Added support for DragonFly BSD, SCO OpenServer 5, SCO OpenServer 6 and
SCO UnixWare 7 to MCE::Util::get_ncpu.
* Also, validated MCE on FreeBSD, NetBSD, OpenBSD, PC-BSD and JabirOS.
No further changes to MCE::Util::get_ncpu.
1.516 Fri Oct 03 02:00:00 EST 2014
[BUG FIXES]
* Updated IPC for better stability across multiple environments. The fix
addresses an issue on Windows where sockets fail to respond after a
period of inactivity; i.e. 4 minutes. Added George Bouras to CREDITS
for reporting the issue.
* Tip for folks developing on Windows:
Open an explorer window and go to C:\Users\<user>\AppData\Local\Temp.
Right-click on the 'mce' folder and create a shortcut on the desktop.
Although MCE removes its temp files automatically, it is possible for
files to remain from a failing app. Go inside the 'mce' folder and
press ctrl-a to select all files. Then press shift-delete;
(macbook[pro] folks: fn-shift-delete).
[ENHANCEMENTS]
* Corrections to documentation.
1.515 Thu Jul 24 23:30:00 EST 2014
[BUG FIXES]
* bin/mce_grep: ${^CHILD_ERROR_NATIVE} is not defined in Perl 5.8.x.
Changed to $?.
[ENHANCEMENTS]
* Tweaked bin/mce_grep. Compute chunk_level => 'auto' to use 'file' when
reading STDIN. Set chunk_size to 8M when not specified (4M previously).
* Added the following names to CREDITS.
Stephan Kulow ; for making the OpenSUSE MCE package
Henry Lu ; for listening while I chatted away about MCE
Jillian Rowe ; for reporting IO::File failing as input_data
Sylvia Roy ; for driving while I worked in the passenger seat
Tom Sasser ; for reporting bin/mce_grep failing with Perl 5.8.x
Florian Schlichting ; for making the Debian MCE package
[NEW FEATURES]
* Added support for IO::File handles as valid input_data including
IO:Uncompress:Gunzip.
Modified SYNTAX section for INPUT_DATA in MCE::Core.pod.
input_data => $fh, ## new IO::File "file", "r"
input_data => $fh, ## new IO::Uncompress::Gunzip "file.gz"
1.514 Thu Jun 05 09:00:00 EST 2014
[BUG FIXES]
* Fixed typo in MCE::Step POD (RT#95250) (Florian Schlichting).
* Updated MCE::Util's get_ncpu function for AIX (Dana Jacobsen).
* Do not send a KILL signal after receiving a SIGPIPE.
* Fixed issue with Makefile.PL on defining the minimum Perl version.
* Use Scalar::Util (looks_like_number) for validation logic. The regex
previously were insufficient for large numbers containing scientific
notation. This impacted the sequence option in MCE.
[ENHANCEMENTS]
* The mce-sandbox demo has been released on GitHub demonstrating
Perl + MCE + Inline::C. The theme is Prime Numbers. This journey
which began 2 years ago has been completed. There is also the
mce-sort exercise demonstrating Perl + MCE + External C.
https://github.com/marioroy/mce-sandbox (is thread-safe)
https://github.com/marioroy/mce-sort (not thread-safe)
* A hard decision had to made for MCE, particularly Perl under Windows
excluding Cygwin. MCE will now load the 'threads' module automatically
for Windows only. Folks may specify use_threads => 0 if threads is
not desired.
The reason for this is from seeing Math::Prime::Util crashing once
workers exit. The same is true without MCE and forking a child process.
Threads does not exhibit this behavior. It is a hard problem to solve.
Why not default to threads for Windows since forking is emulated.
1.513 Sat Apr 19 20:30:00 EST 2014
[BUG FIXES]
* Added fix for bug RT#94869 -- crash when restarting workers with 9+
workers. Updated the perldoc for restart_worker in MCE::Core.pod.
[ENHANCEMENTS]
* Replaced $self with $mce in MCE::/Core.pod to be consistent with
examples described in MCE::Examples.pod.
1.512 Fri Apr 18 21:00:00 EST 2014
[BUG FIXES]
* Signal-handling update for MCE::Signal. Passing the -setpgrp option is not
necessary, even with Daemon::Control. Piping data into and out is better
supported with this release (\*STDIN). This resolves bug RT#94706.
cat infile | mce_script | head
mce_script < infile | head
Added Shawn Halpenny to the CREDITS file.
* The utf8.pl example now runs under the BSD 9.0 environment. This was
failing due to $^H{charnames} is not defined error. Removed the constant
from the list of unicode characters inside the script.
[ENHANCEMENTS]
* Added examples/pipe1.pl and pipe2.pl. These process STDIN or FILE in
parallel. Processing is via Perl for pipe1.pl, whereas an external
command for pipe2.pl.
1.511 Fri Apr 04 22:30:00 EST 2014
[BUG FIXES]
* Added "use bytes;" in several files to have length() return physical
bytes, not logical characters. MCE is now UTF-8 safe when passing scalar
data between workers and the manager process.
Added Marcus Smith to the CREDITS file for reporting this bug.
[ENHANCEMENTS]
* Added examples/utf8.pl
1.510 Thu Mar 27 10:00:00 EST 2014
[ENHANCEMENTS]
* The user_begin and user_end functions now receive 3 arguments.
my ($mce, $task_id, $task_name) = @_;
* Pass the chunk_size value when calling the iterator function. Added a
DBI example under "SYNTAX for INPUT_DATA" in MCE::Core.pod.
* Store the last scalar reference to not have workers re-spawn
unnecessarily when input_data => \$same_scalar_ref.
[NEW FEATURES]
* New parallel_io option to further enhance slurp IO when specifying
use_slurpio => 1 and chunk_size is greater than 8192. Try with
chunk_size => '300k' or chunk_size => '2m'.
The parallel_io option is beneficial when reading from fast storage.
However, possibly not recommended if running MCE on many compute nodes
and having workers read various input_data from shared storage.
Enable parallel_io only if it makes sence and without impacting the
environment such as nfsd. Because use_slurpio => 1, parallel_io => 1,
and chunk_size => '2560k' is truly parallel.
1.509 Sat Feb 03 05:30:00 EST 2014
[BUG FIXES]
* Fixed an issue with all the models (Flow, Grep, Loop, Map, Step, and
Stream) ending immediately on subsequent runs when input_data is
specified through the init method.
* Things have finally settled down with all the models.
[ENHANCEMENTS]
* Refactored the fix applied in 1.508 for addressing bug #92627.
* MCE, for the most part, has been completed. This release touches up on
many examples under the 'examples' directory.
* Brought the MCE::Examples Perl documentation up to date.
Inlined MCE::Loop snippets under the chunking sections.
Added a new section GLOBALLY SCOPED VARIABLES AND MCE MODELS.
Added a new section MONTE CARLO SIMULATION.
1.508 Sat Feb 01 04:00:00 EST 2014
[BUG FIXES]
* Applied fix for bug #92627 submitted by Philip Mabon. MCE scripts may
exit with a non zero error code due to leaving workers up thereby
having MCE performing the shut down in its END block.
For folks running on an older release, the workaround is to shut down
workers prior to exiting.
1.507 Fri Jan 30 23:00:00 EST 2014
[ENHANCEMENTS]
* Added clarity around preserving output order for several models;
MCE::Flow, MCE::Loop, and MCE::Step.
* Inlined comments in the code for overriding MCE options when using
an anonymous hash for the first argument; MCE::Flow, MCE::Step, and
MCE::Stream.
* Removed unnecessary code due to one not able to pass MCE options other
than through the init method for MCE::Loop, MCE::Grep, and MCE::Map.
* Added to doc a use-case example searching a large file with mce_grep_f.
The emphasis is comparing the memory consumption against the native
grep function as well as time to complete.
1.506 Fri Jan 30 03:00:00 EST 2014
[BUG FIXES]
* Fixed issue with all 5 models croaking for an internal '_file' option.
This has been broken since the 1.502 release. Methods impacted were:
mce_flow_f, mce_grep_f, mce_loop_f, mce_map_f, and mce_stream_f
* Specifying a different value for gather on subsequent runs is now taking
effect for when workers persist after running. The models impacted are
MCE::Flow and MCE::Loop.
* Updated _parse_chunk_size in MCE::Util to compute chunk_size correctly
for new edge cases. One may call either mce_grep or mce_grep_f for a
GLOB or scalar reference as input data. The same also applies for the
other models; mce_flow, mce_loop, mce_map and mce_stream.
[ENHANCEMENTS]
* Updates to MCE::Grep and MCE::Map. The logic is now aware of wantarray
for faster processing when storing to a scalar value. In addition, the
use_slurpio option is enabled for efficient IO when processing large
files.
* The task_end option can now be specified for MCE::Stream, although being
used internally.
* The interval option can take a decimal number. Previously, this
wanted a hash reference with up to 3 key/value pairs. Most often, all
one needed was delay and not max_nodes or node_id.
interval => { delay => 0.05 } ## Choose either format for specifying
interval => 0.05 ## delay in 1.506 and above. MCE will
## translate 0.05 to { delay => 0.05 }
## automatically.
* Slight optimization to get_ncpu (public method) in MCE::Util.
* Tweak to _parse_chunk_size (private method) in MCE::Util.
[NEW FEATURES]
* Added a new model MCE::Step for transparent use of MCE::Queue when passing
data among sub-tasks. MCE::Step is basically a spin off from MCE::Flow
with a touch of MCE::Stream.
This new model is crazy :)
1.505 Thu Jan 21 01:30:00 EST 2014
[BUG FIXES]
* Delete $self->{input_data} inside the worker immediately during spawning
(if input_data is an ARRAY, GLOB, or Iterator reference).
* Reverted the logic for the RS (record separator) option from the 1.4x
code base. This is now working as expected.
* Specifying chunk_size => 'auto' via the 'init' method (all 5 models)
causes MCE validation to croak due to 'auto' being carried over to the
core API. Note that only the 5 models are allowed 'auto' for chunk_size
as of this time.
* Update to the MCE->do method. Previously, an undef sent back from the
callback function ended up as a blank value "". Both "" and undef are
now properly captured and sent back to the worker process.
* Added an if statement inside the 'abort' method.
* Removed an old unlink _store.db statement (old code left behind).
* Removed an unused _next variable inside Request.pm.
[ENHANCEMENTS]
* Enhanced egrep.pl to handle additional options including recursion.
This script now supports many egrep options [ceHhiLlmnqRrsv].
* IO performance for examples/cat.pl was improved. The real focus here is
demonstrating output order.
* The chunk_size option can take a suffix; K (Kilobytes) or M (Megabytes).
* The following examples except --max-workers=NUM --chunk-size=NUM options.
cat.pl, egrep.pl, findnull.pl, scaling_pings.pl, and wc.pl
[NEW FEATURES]
* The input_data option can now receive an iterator reference. Added a
new example iterator.pl for demonstraton. In addition, there are
several examples listed under a new section "SYNTAX for INPUT_DATA"
under MCE::Core.pod.
* Added a new demo script; bin/mce_grep. This is a wrapper script for the
grep binary. This script supports agrep, grep, egrep, fgrep, & tre-agrep.
Simply create a link to mce_grep or make a copy. Both Windows and Cygwin
are supported as well. Recursion works for all binaries including agrep
(-R, -r options).
ln mce_grep mce_agrep
ln mce_grep mce_egrep
ln mce_grep mce_fgrep
ln mce_grep mce_tre-agrep
Try with the --lang=C option for faster execution time (-i runs faster).
Try mce_agrep or mce_tre-agrep against very large files. The speedup is
linear and makes good utilization of all available cores on the box.
One may specify the chunking level via the --chunk-level option. For large
files, specify 'file' (chunks file). For many small files, use 'list'.
I have tested against the following GnuWin32 packages found at this URL.
http://gnuwin32.sourceforge.net/packages.html
tre-0.7.5-bin.zip ## Contains agrep.exe (also runs under Cygwin)
grep-2.5.4-bin.zip ## Contains egrep.exe, fgrep.exe, grep.exe
libiconv-1.9.2-1-bin.zip ## These are required for the GnuWin32 binaries.
libintl-0.14.4-bin.zip
pcre-7.0-bin.zip
regex-2.7-bin.zip
Btw, bin/mce_grep is optional and therefore not installed by default
when running make install for the MCE module. Simply copy mce_grep as
mce_agrep.pl for Windows. The '.pl' suffix is optional for other
environments.
1.504 Tue Oct 29 16:00:00 EST 2013
[BUG FIXES]
* MCE::Grep, MCE::Map, and MCE::Stream were failing for mce_grep_s,
mce_map_s, and mce_stream_s when specifying chunk_size => 1.
[ENHANCEMENTS]
* Calibrated 'auto' slightly in MCE::Util::_parse_chunk_size.
1.503 Mon Oct 28 17:00:00 EST 2013
[BUG FIXES]
* The previous release introduced a bug by removing the line closing
STDERR and STDOUT prior to workers exiting. The proper fix intended
is to flush (not close) the handles. Closing them was not the thing
to do in the first place due to possibly needed by the END block or
WARN/DIE handlers initiated from inside the END block.
[ENHANCEMENTS]
* Updated README under examples/tbray and examples/matmult.
* Changed 0.499 to 0.5 inside the yield method.
* Small cosmetic changes otherwise.
1.502 Mon Oct 21 16:00:00 EST 2013
[BUG FIXES]
* The END block for all models will return immediately when called by a
worker thread or process. Removed the line closing STDERR and STDOUT
prior to workers exiting. It turns out that workers were calling
shutdown which is not allowed. This impacted MCE::Flow, MCE::Grep,
MCE::Loop, MCE::Map, and MCE::Stream.
* An update was applied to DESTROY in MCE::Queue to address an edge
case during additional testing across several environments.
* Added documentation describing the core methods in MCE. I had moved the
missing sections to another file some time back and totally forgotten
about them when finalizing on MCE::Core.pod for the 1.5 release.
[ENHANCEMENTS]
* All models will croak when specifying an invalid MCE option.
[NEW FEATURES]
* One can specify a hash reference for the gather option. Updated
documentation describing the gather option and the gather method.
1.501 Wed Oct 16 01:20:00 EST 2013
[BUG FIXES]
* Addressed an issue with the worker signal "die" handler (#89538). Eval'd
code should not raise an exception causing the app to die. This is working
as expected for workers spawned as children.
For threads, a die called explicitly inside an eval block will cause the
worker to exit (same as previously). However, an eval 'use MissingModule'
will no longer cause the thread to die.
[ENHANCEMENTS]
* Changelog section uses a new format beginning with this release.
[NEW FEATURES]
* Added examples/flow_model.pl for demonstrating MCE::Flow, MCE::Queue, and
MCE->gather.
1.500 Thu Oct 10 01:00:00 EST 2013
* MCE 1.5 is backwards compatible with 1.4 and below.
* The documentation, previously MCE.pod, moved to MCE::Core.pod with the
examples section placed in MCE::Examples.pod. The MCE.pod file serves
as an index page for the various documentation.
* IPC has been enhanced with 8 data channels. Many operations run 3x when
compared with MCE 1.4.
* Five models: MCE::Flow, MCE::Grep, MCE::Loop, MCE::Map, and MCE::Stream.
* Hybrid queues via MCE::Queue allowing for normal and priority queues.
* MCE::Subs for exporting functions prefixed with mce_; e.g. mce_wid.
* All public methods can be called directly using the package name
and method e.g. MCE->wid, MCE->run.
MCE->new(
max_workers => 'auto',
user_func => sub {
my $wid = MCE->wid;
MCE->sendto("STDOUT", "Hello from $wid\n");
}
);
MCE->run;
* Localize the input scalar $_ prior to calling user_func. Folks can
use $_ for input_data and sequence of numbers. Added section to docs
explaining DEFAULT INPUT SCALAR.
* New options (bounds_only, gather, interval, task_name).
The task_end option can now be specified at the top level.
See docs for use case.
Input_data can be specified inside the first task instead of having
to specify this at the top level. Input_data is ignored when specified
for tasks other than the first task.
user_tasks => [{
input_data => \@list,
...
},{
...
}
* New public methods
chunk_id, gather, freeze, thaw, yield, task_name, print, printf, say
* New example (interval.pl).
* Optimized the egrep.pl and wc.pl examples. These run much faster.
Try these out against large log files. Both examples fly.
* Barrier synchronization update. Two sockets are utilized instead of
2 lock files. This, now works wonderfully under the Cygwin environment.
For threading, the removal of 2 lock files increases the number of
threads allowed from about 1/3rd previously to under 1/2 of ulimit,
e.g. int(ulimit -n / 2 - 20).
* Removed the logic for determining MAX_OPEN_FILES and MAX_USER_PROCS.
MCE no longer has a constraint on max_workers allowed.
* Code re-factor work. Added a private method _validate_runstate called
by various methods. Organized the code slightly such as placement of
methods.
1.415 Mon Jun 17 15:00:00 EST 2013
* Code-refactor in preparation for the upcoming 1.5 release.
This completes the 1.4x branch for a very solid and stable release.
* IPC optimization. Localize $\ and $/ only as needed. Small tweak to
logic when specifying RS (record separator) for input data.
Foreach is very communication intensive. Forseq is less so. Chunking
was made faster by increasing chunk size from 500 to 2500. I've reached
my goal for forseq by reaching 60,000+. A while back, foreach could
barely reach 18,000.
Before and after results with IPC optimization.
(Before) (After)
Parallel::Loops 600 600
$mce->foreach 20,000 21,500 (+ 1,500)
$mce->forseq 55,000 64,000 (+ 9,000)
$mce->forchunk 395,000 450,000 (+ 55,000)
* Updated the process method. Allows sequence to be specified as an
option. Also allow the hash to be specified as the 1st or 2nd
argument.
$mce->process( \@input_array, { options } );
$mce->process( { options }, [ 1..1000 ] );
$mce->process( { sequence => [ 20, 40, 2 ] } );
* The forseq method will now honor chunk_size greater than 1. Updated
docs with use case.
* Added CONST short for EXPORT_CONST, e.g. use MCE CONST => 1;
1.414 Sun Jun 16 13:00:00 EST 2013
* Upstream bug fixes. Mainly chunk_size was not honored when specifying
user_tasks and sequence/chunk_size for the first task. Bug is with
not setting abort_msg correctly.
1.413 Mon Jun 10 02:00:00 EST 2013
* Fixed typo: Changed local @_ to local $@ inside _parse_max_workers
before evaling.
1.412 Sun Jun 09 14:00:00 EST 2013
* Updated the main README file and CHANGES under 1.411 below.
* Added support for multiplication and division as well as
case insentivity to the private _parse_max_workers method.
max_workers => 'auto', ## = MCE::Util::get_ncpu()
max_workers => 'Auto-1', ## = MCE::Util::get_ncpu() - 1
max_workers => 'AUTO + 3', ## = MCE::Util::get_ncpu() + 3
max_workers => 'AUTO * 1.5',
max_workers => 'auto / 1.333',
1.411 Sun Jun 09 02:00:00 EST 2013
* Reverted the change made in MCE-1.410. That degraded slurp IO on
large files. This restores slurp IO performance in MCE.
* Corrected an if statement towards the end of the shutdown method.
The delay was occurring always when it was meant only if MCE was
launched from inside a non-main thread.
* Addressed a race condition with barrier synchronization under the
Cygwin environment. Was not able to reproduce this with other OS's.
This takes barrier synchronization work to 100%. Delay statements
for Cygwin were removed around barrier synchronization.
* Optimized writes to sockets. Arguments passed to print statements
are concatenated as one big string. Removed local $/ = $LF when
reading from the queue socket. It's not required there since using
read and known size. The foreach.pl example (very communication
intensive) can now do 20000 in 1 second on my Macbook Pro. Updated
benchmark results inside foreach.pl, forchunk.pl, and forseq.pl.
* MAX_WORKERS, CHUNK_SIZE, TMP_DIR, FREEZE and THAW can be specified
when loading the module. FREEZE and THAW allows one to choose an
alternative serialization module if preferred for your project.
use Sereal qw(encode_sereal decode_sereal);
use MCE FREEZE => \&encode_sereal, THAW => \&decode_sereal;
* Passing EXPORT_CONST => 1 will export 3 constants: SELF, CHUNK, CID
The "my ($self, $chunk_ref, $chunk_id) = @_" line is not necessary.
$_[SELF], $_[CHUNK], $_[CID] refers to $_[0], $_[1], $_[2].
* The MCE::Util module was created. It contains the get_ncpu function.
This is largely borrowed from Test::Smoke::Util.pm. MCE supports
'auto' when specifying max_workers. Read doc for other use case.
max_workers => 'auto'; ## Same as MCE::Util::get_ncpu()
* Re-factored code in preparation for the upcoming 1.5 release. The
1.5 release will use just 3 socket pairs versus 4. I did not want
to include that change in this release. Therefore, I decided to
take the upstream bug fixes and apply them to the 1.4 base.
Re-factored documentation by making use of the full 78 character
width per line.
* Updated the main README file under the top level dir.
1.410 Tue May 28 23:30:00 EST 2013
* Use threads under MSWin32 for 02* and 03* test scripts.
* Removed sysopen/sysseek/sysread to simplify logic due to negligible
performance gains over open/seek/read.
* Minor updates to documentation.
1.409 Sun May 12 22:45:00 EST 2013
* Croak if user_func is not defined and input_data/sequence is specified.
* Fix barrier synchronization when running multiple tasks via user_tasks.
* Updated Perl documentation for easier reading through cpan/metapan.
* Renamed private method _worker_sequence to _worker_sequence_queue.
1.408 Tue Mar 19 22:00:00 EST 2013
* Minor tweaks here and there to further increase reliability.
* Updated the barrier synchronization logic to not stall. Updated the
perl docs on mixing "sync" with "do" or "sendto" methods.
* Added new "status" method for the manager process.
* Added new arguments for MCE::Signal: -no_kill9, -no_sigmsg
1.407 Thu Mar 14 21:00:00 EST 2013
* This marks a tremendous effort in achieving parity across the board
from Cygwin to Windows and obviously UNIX. MCE now works beautifully
under the Cygwin environment including Perl for Windows.
* Up'ed the maximum workers allowed for both Cygwin and Windows to
56 (forking) and 80 (threading).
* Barrier synchronization requires an extra semaphore file, therefore
the maximum workers allowed under the UNIX environment for threading
decreased. It shouldn't be a problem as there are many threads already.
* Addressed an issue with barrier synchronization under Cygwin.
* Addressed an issue with the die handler for the main worker method
when threading.
* Thank you for hanging in there. It took quite some time to get there.
This is the most stable release thus far for MCE across the board.
1.406 Tue Mar 12 19:00:00 EST 2013
* Added support for barrier synchronization (via new sync method).
Added barrier_sync.pl example.
* Addressed rounding errors with the sequence generator.
The sequence option now follows a bank-teller queuing model when
generating numbers. This applies to task 0 only. Other tasks can
still specify sequence where numbers will be distributed equally
among workers like before.
* Optimized the _worker_request_chunk private method.
* A take 2 on the matrix multiplication examples. This is better
organized with updated README file to include the script running
time in the results.
1.405 Mon Mar 04 19:00:00 EST 2013
* Added strassen_pdl_t.pl in the event folks cannot make use of /dev/shm
used by the strassen_pdl_s.pl example.
* Optimized the 'send' method -- workers process immediately after receiving
data. Updated run times in README for the strassen examples.
* MCE no longer calls setpgrp by default as of MCE 1.405. There is only
one reason to call setpgrp, but many reasons not to. The sole reason
was for MCE to run correctly with Daemon::Control. If needed, one can
pass the option to MCE::Signal qw(-setpgrp).
* Return void in the shutdown method (previously was returning $self).
* Tidy code inside sequence generator.
1.404 Sun Feb 24 13:00:00 EST 2013
* Added sess_dir method
* Completed work with matmult/* examples
Added matmult_pdl_q.pl, removed strassen_pdl_h.pl
Added strassen_pdl_o/p/q/r/s.pl
Added benchmark results from a 32-way box at the end of the readme
* Removed lines setting max limit for files/procs
1.403 Sun Feb 17 15:00:00 EST 2013
* Wrap sub PDL::CLONE_SKIP into a no warnings 'redefine' block
MCE now works with PDL::Parallel::threads without any warnings
* Added missing examples/matmult/matmult_pdl_n.pl to MANIFEST
* Refactored strassen examples, memory consumption reduced by > than 50%
* Added matmult_pdl_o.pl -- uses PDL::Parallel::threads to share matrices
* Added matmult_pdl_p.pl -- matrix b is read from shared memory, not mmap
* Added strassen_pdl_n.pl -- additional improvements to memory reduction
* Added strassen_pdl_h.pl -- shown running with 4 workers (half and half)
* Re-ran matrix multiplication examples and updated results in README file
* Added -no_setpgrp option to MCE::Signal.pm
Ctrl-C does not respond when running /usr/bin/time mce_script.pl
* Added undef $buffer in a couple of places within MCE.pm
* Added David Mertens and Adam Sjøgren to CREDITS
* The 'send' method now checks if sending > total workers after spawning
not before
1.402 Thu Feb 14 07:30:00 EST 2013
* Updated matrix multiplication examples including README
* Added examples/matmult/matmult_pdl_n.pl
1.401 Tue Feb 12 19:00:00 EST 2013
* Added sub PDL::CLONE_SKIP { 1 } to MCE.pm. Running PDL + MCE threads no
longer crashes during exiting.
* Updated matrix multiplication examples. All examples now work under the
Windows environment no matter if threading or forking. Unix is stable as
well if wanting to use PDL + MCE and use_threads => 1 or 0.
* Added benchmark results for 2048x2048, 4096x4096, and 8192x8192 to the
README file under examples/matmult/
* Updated documentation
1.400 Mon Feb 11 07:00:00 EST 2013
* Slight optimization in the _do_callback method
* Added 2 new options: user_args and RS (record separator)
* Added new send method for sending data to workers after spawning and
prior to running
* The sequence option can now take an ARRAY reference
* Updated documentation on new features
* Added matrix multiplication examples
1.306 Sat Jan 05 18:00:00 EST 2013
* Added if statement around setpgrp(0,0). That function is not supported
under Windows.
* Updated logic for removing any remaining MCE session directories inside
MCE::Signal.
1.305 Sat Jan 05 16:00:00 EST 2013
* Added check for $^S to the DIE handler inside the _worker_main method
* Added setpgrp(0,0) to MCE::Signal's BEGIN block
* MCE::Signal points to a _mce_sess_dir hash in the event of a signal,
will remove the sess_dir(s) as well. This is needed when tmp_dir is
specified during instantiation and pointing to another location than
MCE::Signal::tmp_dir.
1.304 Wed Jan 02 22:00:00 EST 2013
* Added Oliver Gorwits to CREDITS for identifying 2 issues
* Direct die to CORE::die inside handler if executing an eval
* Undef $mce_spawned_ref if signal was caught (stop_and_exit)
* Changed INIT to sub import in MCE.pm
1.303 Tue Jan 01 20:00:00 EST 2013
* Bump version -- MCE.pm VERSION now matches with META.yml
* Sorted forchunk, foreach, forseq methods inside MCE.pm
* Modified if statement in run method
* Task ID is never undef, therefore removed check inside
restart_worker and worker_do methods
* Added 2 package variables to MCE::Signal
$display_die_with_localtime and $display_warn_with_localtime
* Completed updates to documentation
* Update to forseq.pl and seq_demo.pl examples
1.302 Tue Jan 01 07:30:00 EST 2013
* Fixed validation logic for sequence
* Updated the sequence generator -- now supports chunking
* Updated seq_demo.pl example to demo user_tasks, sequence, and
chunk_size applied together
* Documentation updates
1.301 Mon Dec 31 15:00:00 EST 2012
* Emphasis on documentation -- better flow plus additional clarity
* Minor updates to sequence option validation
* Minor updates to included examples
1.300 Mon Dec 31 06:00:00 EST 2012
* New methods...: chunk_size, restart_worker, task_id, task_wid, tmp_dir
* New options...: on_post_exit, on_post_run, sequence
* New examples..: forseq.pl, seq_demo.pl
* Overhaul to exit method
Workers can exit or die without impacting the manager process
* Enabled executable bit for test files
* Removed localtime output in die and warn handlers
* All 3 delay options are consistent whether user_tasks is specified
* Removed logic around total_ended count -- replaced with new exit logic
* Code refactoring plus documentation updates
* Added LICENSE file
1.201 Fri Dec 21 00:00:00 EST 2012
* Added MCE.pod -- moved documentation from MCE.pm to pod file
* Added missing use strict/warnings to test scripts
* Default to 1 for chunk_size and max_workers if not specified
* Test::More is not a requirement to run MCE, only for building
* Changed the format for the change log file
1.200 Thu Dec 20 00:00:00 EST 2012
* Added new user_tasks option
* Added space between method name and left-paren for header lines in POD
* Remove not-needed BSD::Resource and forks inside BEGIN/INIT blocks
1.106 Wed Dec 19 05:00:00 EST 2012
* Added t/pod-coverage.t
* Big overhaul of the MCE documentation -- all methods are documented
* Croak if method suited for a MCE worker is called by the main MCE process
* Croak if method suited for the main MCE process is called by a MCE worker
* Updated Makefile.PL to declare the minimum Perl version
1.105 Sun Dec 16 23:00:00 EST 2012
* Completed code re-factoring
* Added t/pod.t
1.104 Sun Nov 25 17:00:00 EST 2012
* Added 1 new example to MCE's Perl documentation
* Use module::method name versus constant symbol when calling _croak
* Croak if session directory is not writeable inside MCE::spawn
* Renamed _mce_id to _mce_sid (meant to be spawn id actually)
* Re-calibrated maximum workers allowed
1.103 Fri Nov 23 13:00:00 EST 2012
* Added writeable check on /dev/shm
* Croak if tmp dir is not writeable inside MCE::Signal::import
1.102 Thu Nov 22 13:00:00 EST 2012
* Woohoot !!! MCE now passes with Perl 5.17.x
* Added Copying file -- same as in Perl
1.101 Wed Nov 21 16:00:00 EST 2012
* Shifted white space to the left for code blocks inside documentation
1.100 Wed Nov 21 10:00:00 EST 2012
* Completed optimization and shakeout for MCE's existing API
* File handles are cached when calling sendto and appending to a file
* The sendto method now supports multiple arguments -- see perldoc
* Added new option: flush_file
1.008 Sat Nov 17 23:00:00 EST 2012
* Updates to __DIE__ and __WARN__ handling in MCE. These address the
unreferenced scalars seen in packaging logs at activestate.com for
Perl under Windows: https://code.activestate.com/ppm/MCE/
* Update t/01_load_signal_arg.t -- added check for $ENV{TEMP}
This fixes issue seen under Cygwin
1.007 Thu Nov 15 21:30:00 EST 2012
* At last, the "Voila" release :)
* Small change to __DIE__ and __WARN__ signal handling for spawn method
1.006 Thu Nov 15 03:30:00 EST 2012
* Added description section to MCE::Signal's Perl doc
* Do not set trap on __DIE__ and __WARN__ inside MCE::Signal
* Localized __DIE__ and __WARN__ handlers inside MCE instead
* Clarify the use of threads in documentation
1.005 Tue Nov 13 06:11:00 EST 2012
* Removed underscore from package globals in MCE::Signal
* Optimized _worker_read_handle method in MCE
* Updated files under examples/tbray/
1.004 Mon Nov 12 01:50:00 EST 2012
* Updated examples/mce_usage.readme
* Updated examples/wide_finder.pl
* Added examples/tbray/README
* Added examples/tbray/tbray_baseline1.pl
* Added examples/tbray/tbray_baseline2.pl
* Added examples/tbray/wf_mce1.pl
* Added examples/tbray/wf_mce2.pl
* Added examples/tbray/wf_mce3.pl (../wide_finder.pl moved here)
* Added examples/tbray/wf_mmap.pl
1.003 Sat Nov 10 12:55:00 EST 2012
* Updated README
* Updated images/06_Shared_Sockets.gif
* Updated images/10_Scaling_Pings.gif
* Added images/11_SNMP_Collection.gif
* Minor updates to MCE::Signal
1.002 Thu Nov 08 01:13:10 EST 2012
* Renamed continue method to next
1.001 Wed Nov 07 23:58:20 EST 2012
* Added perl-MCE.spec to trunk
https://code.google.com/p/many-core-engine-perl/source/browse/trunk/
* Added CREDITS
* Added 3 new methods to MCE.pm: continue, last, and exit
* Both foreach & forchunk now call run(1, {...}) to auto-shutdown workers
1.000 Mon Nov 05 10:00:00 EST 2012
* First release