The following scripts test the various components that make up GNU parallel.

They are based on examples in the GNU parallel man pages.

Note that the last one for sql has been obstificated to not disclose the

testers username and password or the name of the database.

There are lots more examples in the GNU parallel man page that can be used

to provide further testing if needed.

----

$ cat niceload-test.sh

#!/bin/sh

PATH=/usr/bin

echo "Testing niceload"

echo

echo "EXAMPLE: See niceload in action"

echo "     In terminal 1 run: top"

echo

echo "     In terminal 2 run:"

echo

echo "     niceload perl -e '$|=1;do{$l==$r or print \".\";"

echo "     $l=$r}until(($r=time-$^T)>50)'"

echo

echo "     This will print a '.' every second for 50 seconds and eat a"

echo "     lot of CPU. When the load rises to 1.0 the process is"

echo "     suspended."

echo

niceload -q perl -e '$|=1;do{$l==$r or print "."; $l=$r}until(($r=time-$^T)>50)'

----

$ cat parallel-test.sh

#!/bin/sh

PATH=/usr/bin:/usr/sbin

echo Testing parallel...

echo

echo EXAMPLE: Composed commands

echo     A job can consist of several commands. This will print the

echo     number of files in each directory:

echo

rm -rf test-dir

mkdir -p test-dir/a test-dir/b test-dir/c test-dir/d

(cd test-dir/a; touch 1 2 3 4 5 6 7 8 9 10 11 12)

(cd test-dir/b; touch 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20)

(cd test-dir/c; touch 1 2 3 4 5 6 7 8)

(cd test-dir/d; touch 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15)

(cd test-dir; ls | parallel 'echo -n {}" "; ls {}|wc -l')

echo

echo "    To put the output in a file called <name>.dir:"

echo

(cd test-dir; ls | parallel '(echo -n {}" "; ls {}|wc -l) > {}.dir')

cat test-dir/a.dir

cat test-dir/b.dir

cat test-dir/c.dir

cat test-dir/d.dir

echo

echo EXAMPLE: Group output lines

echo     When running jobs that output data, you often do not want

echo     the output of multiple jobs to run together. GNU parallel

echo     defaults to grouping the output of each job, so the output

echo     is printed when the job finishes. If you want the output to

echo     be printed while the job is running you can use -u.

echo

echo     Compare the output of:

echo

parallel traceroute ::: foss.org.my debian.org freenetproject.org

echo

echo     to the output of:

echo

parallel -u traceroute ::: foss.org.my debian.org freenetproject.org

echo

echo EXAMPLE: Keep order of output same as order of input

echo      Normally the output of a job will be printed as soon as it

echo      completes. Sometimes you want the order of the output to

echo      remain the same as the order of the input. This is often

echo      important, if the output is used as input for another

echo      system. -k will make sure the order of output will be in the

echo      same order as input even if later jobs end before earlier

echo      jobs.

echo

echo      Append a string to every line in a text file:

echo      If you remove -k some of the lines may come out in the wrong

echo      order.

echo

echo BEFORE:

echo

cat textfile

echo

echo AFTER:

echo

cat textfile | parallel -k echo {} append_string

----

$ cat parcat-test.sh

#!/bin/sh

PATH=/usr/bin

echo "Testing parcat"

echo

echo "mkfifo slot-{1..5}-digit-{0..9}"

echo "parallel -j5 'seq 100000 | grep {} > slot-{%}-digit-{}' ::: {0..9} &"

echo "parallel parcat slot-{1..5}-digit-{} '>' digit-{} ::: {0..9}"

echo

rm -f slot* digit*

mkfifo slot-{1..5}-digit-{0..9}

parallel -j5 'seq 100000 | grep {} > slot-{%}-digit-{}' ::: {0..9} &

parallel parcat slot-{1..5}-digit-{} '>' digit-{} ::: {0..9}

----

$ cat sem-test.sh

#!/bin/sh

PATH=/usr/bin:/usr/sbin:/usr/perl5/5.22/bin

echo Testing sem

echo

echo EXAMPLE: Gzipping *.log

echo      Run one gzip process per CPU core. Block until a CPU core

echo      becomes available.

echo

rm -rf test-dir

mkdir test-dir

mkfile -v 150k test-dir/a.log

mkfile -v 150k test-dir/b.log

mkfile -v 150k test-dir/c.log

mkfile -v 150k test-dir/d.log

mkfile -v 150k test-dir/e.log

mkfile -v 150k test-dir/f.log

mkfile -v 150k test-dir/g.log

mkfile -v 150k test-dir/h.log

mkfile -v 150k test-dir/i.log

mkfile -v 150k test-dir/j.log

mkfile -v 150k test-dir/k.log

mkfile -v 150k test-dir/l.log

mkfile -v 150k test-dir/m.log

mkfile -v 150k test-dir/n.log

mkfile -v 150k test-dir/o.log

mkfile -v 150k test-dir/p.log

mkfile -v 150k test-dir/q.log

mkfile -v 150k test-dir/r.log

for i in `ls test-dir/*.log` ; do

  echo $i

  sem -j+0 gzip $i ";" echo done

done

sem --wait

echo

echo "EXAMPLE: Protecting pod2html from itself"

echo "    pod2html creates two files: pod2htmd.tmp and pod2htmi.tmp"

echo "    which it does not clean up. It uses these two files for a"

echo "    short time. But if you run multiple pod2html in parallel"

echo "    (e.g. in a Makefile with make -j) you need to protect"

echo "    pod2html from running twice at the same time. sem running as"

echo "    a mutex will do just that:"

echo

#cp .../components/perl512/perl-5.12.4/Porting/pumpkin.pod .

sem --fg --id pod2html pod2html pumpkin.pod > pumpkin.html

sem --fg --id pod2html rm -f pod2htmd.tmp pod2htmi.tmp

----

$ cat sql-test.sh

#!/bin/sh

PATH=/usr/bin:/usr/sbin:/usr/perl5/5.22/bin

echo Testing sql

echo

echo EXAMPLE: Attaching to the BugDB database.

echo

sql oracle://<USERNAME>:<PASSWORD>@<DATABASE NAME>

----