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

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

     3 

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

     5 testers username and password or the name of the database.

     6 

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

     8 to provide further testing if needed.

     9 

    10 ----

    11 

    12 $ cat niceload-test.sh

    13 #!/bin/sh

    14 PATH=/usr/bin

    15 echo "Testing niceload"

    16 echo

    17 echo "EXAMPLE: See niceload in action"

    18 echo "     In terminal 1 run: top"

    19 echo

    20 echo "     In terminal 2 run:"

    21 echo

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

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

    24 echo

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

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

    27 echo "     suspended."

    28 echo

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

    30 

    31 ----

    32 

    33 $ cat parallel-test.sh

    34 #!/bin/sh

    35 PATH=/usr/bin:/usr/sbin

    36 echo Testing parallel...

    37 echo

    38 echo EXAMPLE: Composed commands

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

    40 echo     number of files in each directory:

    41 echo

    42 rm -rf test-dir

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

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

    45 (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)

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

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

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

    49 echo

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

    51 echo

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

    53 cat test-dir/a.dir

    54 cat test-dir/b.dir

    55 cat test-dir/c.dir

    56 cat test-dir/d.dir

    57 echo

    58 echo EXAMPLE: Group output lines

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

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

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

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

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

    64 echo

    65 echo     Compare the output of:

    66 echo

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

    68 echo

    69 echo     to the output of:

    70 echo

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

    72 echo

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

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

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

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

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

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

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

    80 echo      jobs.

    81 echo

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

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

    84 echo      order.

    85 echo

    86 echo BEFORE:

    87 echo

    88 cat textfile

    89 echo

    90 echo AFTER:

    91 echo

    92 cat textfile | parallel -k echo {} append_string

    93 

    94 ----

    95 

    96 $ cat parcat-test.sh

    97 #!/bin/sh

    98 PATH=/usr/bin

    99 echo "Testing parcat"

   100 echo

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

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

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

   104 echo

   105 rm -f slot* digit*

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

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

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

   109 

   110 ----

   111 

   112 $ cat sem-test.sh

   113 #!/bin/sh

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

   115 echo Testing sem

   116 echo

   117 echo EXAMPLE: Gzipping *.log

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

   119 echo      becomes available.

   120 echo

   121 rm -rf test-dir

   122 mkdir test-dir

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

   142   echo $i

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

   144 done

   145 sem --wait

   146 echo

   147 echo "EXAMPLE: Protecting pod2html from itself"

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

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

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

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

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

   153 echo "    a mutex will do just that:"

   154 echo

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

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

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

   158 

   159 ----

   160 

   161 $ cat sql-test.sh

   162 #!/bin/sh

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

   164 echo Testing sql

   165 echo

   166 echo EXAMPLE: Attaching to the BugDB database.

   167 echo

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

   169 

   170 ----