SYNOPSIS

dnetc [-config] [-flush] [-fetch] [-update] [-benchmark [prj]] [-benchmark2 [prj]] [-bench [prj [cn]]] [-test [prj [cn]]] [-stress [prj [cn]]] [-cpuinfo] [-restart] [-shutdown] [-pause] [-unpause] [-install [...]] [-uninstall] [-import <filename>] [-forceunlock <fn>] [-help] [-version] [-ini <filename>] [-e <address>] [-nodisk] [-n <count>] [-runbuffers] [-frequent] [-inbase <fname>] [-outbase <fname>] [-ckpoint <fname>] [-bin <prj> <n>] [-btime [prj] <n>] [-runoffline] [-runonline] [-nettimeout <secs>] [-a <address>] [-p <port>] [-nofallback] [-u <method>] [-ha <address>] [-hp <port>] [-lurk] [-lurkonly] [-interfaces <list>] [-c [prj] <n>] [-numcpu <n>] [-devicenum <n>] [-priority <0-9>] [-l <filename>] [-smtplen <len>] [-smtpsrvr <host>] [-smtpport <port>] [-smtpfrom <id>] [-smtpdest <id>] [-h <hours[:min]>] [-until <HH:MM>] [-noexitfilecheck] [-pausefile <fn>] [-exitfile <fn>] [-multiok[=|:][0|1]] [-percentoff] [-quiet/-hide] [-noquiet]

DESCRIPTION

dnetc is a distributed computing client that coordinates with servers operated by distributed.net to cooperate with other network-connected computers to work on a common task. It communicates over public networks and processes work assigned by the distributed.net keyservers. It is designed to run in idle time so as to not impact the normal operation of the computer.

INSTALLATION

Since you are already reading this, I assume you know how to unpack an archive into a directory of your choice. :)

Now, simply fire up the client...

If you have never run the client before, it will initiate the menu-driven configuration. Save and quit when done, the configuration file will be saved in the same directory as the client. Now, simply restart the client. From that point on it will use the saved configuration.

The configuration options are fairly self-explanatory and can be run at any time by starting the client with the '-config' option. A list of command line options is listed below.

OPTIONS

In addition to the conventional command line passed to the client from a shell, options may also be passed to the client using either or both of the following methods:

-      using the dnetc_opt= environment variable.

       If set, this is parsed before the normal command line.

-      using the //etc/distributed-net.conf

       command files. If found, these are parsed after the normal

       command line.

"Mode commands" (see below) cannot be executed using these methods, and there is no run-time display of modified settings (unless the settings are also modified using the conventional command line, in which case the last effective change is displayed).

Mode commands: (the client will execute the option and then exit)

-config

start the configuration menu

-flush

flush all output buffers

-fetch

fill all input buffers

-update

fetch + flush

-benchmark [prj]

16-20 sec speed check [optional: only project prj]

-benchmark2 [prj]

half (8-10 sec) and slightly inaccurate -benchmark

-bench [prj [cn]

-benchmark all cores [optional: only project prj]

[optional: only core cn, must be used with prj]

-test [prj [cn]

tests for core errors [optional: only project prj]

[optional: only core cn, must be used with prj]

-stress [prj [cn]

more thorough and longer running variant of -test

-cpuinfo

print information about the detected cpu(s)

-restart

restart all active clients

-shutdown

gracefully shut down all active clients

-pause

pause all active clients

-unpause

unpause all active clients

-install [...]

install the client in /etc[/rc.d]/init.d/

all [...options...] that follow '-install' serve

as parameters for the installed client.

-uninstall

remove the client from /etc[/rc.d]/init.d/

-import <filename>

import packets from file <filename> into client buffers

-forceunlock <fn>

unlock buffer file <fn>

-help

display this text

-version

print version information

Project and buffer related options:

-ini <filename>

override default name of INI file

-e <address>

the email id by which you are known to distributed.net

-nodisk

don't use disk buffer files

-n <count>

packets to complete. -1 forces exit when buffer is empty.

-runbuffers

set -n == -1 (exit when buffers are empty)

-frequent

frequently check if buffers need topping-up

-inbase <fname>

input buffer basename (ie without 'extension'/suffix)

-outbase <fname>

output buffer basename (ie without 'extension'/suffix)

-ckpoint <fname>

set the name of the checkpoint file

-bin <prj> <n>

set fetch buffer threshold to <n> work units

-btime [prj <n> ]

set fetch time threshold to <n> hours

If not specified, project name <prj> defaults to RC5

Network update related options:

-runoffline

disable network access

-runonline

enable network access

-nettimeout <secs>

set the network timeout. Use -1 to force blocking mode

-a <address>

keyserver name or IP address

-p <port>

keyserver port number

-nofallback

don't fallback to a distributed.net keyserver

-u <method>

use this UUE/HTTP encoding method (see -config)

-ha <address>

http/socks proxy name or IP address

-hp <port>

http/socks proxy port

-lurk

automatically detect modem connections

-lurkonly

perform buffer updates only when a connection is detected

-interfaces <list>

limit the interfaces to monitor for online/offline status

Performance related options:

-c [prj <n> ]

core number (run -config for a list of valid core numbers)

project name 'prj' defaults to RC5

-numcpu <n>

run <n> threads/run on <n> cpus. 0 forces single-threading.

-devicenum <n>

run on device <n> only.

-priority <0-9>

scheduling priority from 0 (lowest/idle) to 9 (normal/user)

Logging options:

-l <filename>

name of the log file

-smtplen <len>

max size (in bytes) of a mail message (0 means no mail)

-smtpsrvr <host>

name or IP address of mail (SMTP) server

-smtpport <port>

mail (SMTP) server port number

-smtpfrom <id>

who the client should say is sending the message

-smtpdest <id>

who the client should send mail to

Miscellaneous runtime options:

-h <hours[:min]>

time limit in hours

-until <HH:MM>

quit at HH:MM (eg 07:30)

-noexitfilecheck

override .ini exit flagfile setting

-pausefile <fn>

name of file that causes the client to pause

-exitfile <fn>

name of file that causes the client to exit

-multiok[=|:][0|1]

allow/disallow multiple instances of the client to run

The default is 'allow' for all platforms but Windows.

-percentoff

don't display work completion as a running percentage

-quiet/-hide

suppress screen output (== detach for some clients)

-noquiet

don't suppress screen output (override ini quiet setting)

FILES

/etc/distributed-net.conf

Additional sources of command line options (parsed last)

AUTHOR

distributed.net http://www.distributed.net/