1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
|
Run the test after compiling adios in the trunk. I.e. after "make"
1. Build the codes necessary for the tests and run sequential tests, i.e. do
a "make check" in adios trunk.
2. Run test.sh script in an interactive-job environment to run the parallel tests.
Define the command for running parallel programs like mpirun -np or aprun -q -n
Run the script in the job directory with full path.
Usage: <path>/`basename $0` [-m runcmd] [-n "-np"] [-p procs] [-h] [-v] [-k]
[pattern [pattern2 pattern3 ...]]
<path> is used to find all the test codes, inputs and reference outputs.
codes and inputs will be copied to the job directory
pattern if given, only those tests are executed, that match
'tests/*<pattern>*.sh'; otherwise all tests are executed.
e.g. arguments 1[0-3] attr global will execute all tests that
have any of the strings 10,11,12,13,attr or global in their name
(and end with .sh)
OPTIONS
-m runcmd command to start an MPI program. Default is "aprun -q" if aprun
is found, otherwise "mpirun"
-n runopt option to runcmd to specify number of cores.
Default is "-n" for aprun, "-np" for mpirun
-p procs Run only those tests that use less up to 'procs' processes.
Default: 128
-k Do not remove logs and work dir of successful tests.
-v Verbose mode
-h Print this help.
Each test will have an ./work.<testname> directory to store anything needed.
The test's stdout and stderr is redirected into ./log.<testname> file.
E.g.
$ qsub -I -lwalltime=1:00:00 -lsize=12
$ cd /tmp/work/pnorbert/test/adios/suite
$ /ccs/proj/e2e/pnorbert/ADIOS/ADIOS/trunk/tests/suite/test.sh -p 12
Settings:
Test source directory: /ccs/proj/e2e/pnorbert/ADIOS/ADIOS/trunk/tests/suite
Run command: aprun -q
Run command np option: -n
Max. processes to use: 12
Keep successful tests' output: no
Run tests
Test 01_scalars ... OK
Test 02_attributes ... OK
Test 03_arrays ... OK
...
$
|
