View Source

This section describes how to set up and run the Lehigh University Benchmark (LUBM) \[12\] using the scripts and configuration files included with the OWLIM-Lite distribution. Running the tests and the benchmarks may require minor modifications to these scripts and configuration files, as discussed in this section.

{toc}


h1. Configuring OWLIM-Lite and Sesame

Running the benchmark (as well as the getting started application), can require modification of the script setvars (.cmd or .sh) in OWLIM-Lite's main folder. The most important setting is the specification of the Java virtual machine through the JAVA_HOME environment variable.

The performance of the benchmark can be tuned to the particular hardware on which it is running with modification of the repository specification file named {{lubm.ttl}}, which can be found in the {{lubm}} sub-directory of OWLIM-Lite's main directory.

h1. Running the Benchmark

To run the Lehigh University Benchmark (LUBM), it is necessary to generate the test file-set beforehand. This can be done using the {{lubm-generate}} (.cmd or .sh, respectively for Windows or Linux) script, which is part of the distribution's {{lubm}} sub-folder. The distribution includes a pre-built library of the [benchmark's source code|http://swat.cse.lehigh.edu/projects/lubm/index.htm]. The pre-built {{lubm.jar}} library is located in the {{ext}} distribution folder and also includes the wrapper classes that the benchmark's code-base uses in order to run against a Sesame repository (configured with OWLIM-Lite).
The {{lubm-generate}} script (.cmd or .sh) accepts a single numeric argument as the target number of universities and creates a sub-folder for the test run with the name {{univer}} and the number of universities appended, e.g. {{univer1000}}. This is the folder in which it places the generated OWL files. Once the file-set is generated, the {{lubm.config}} file should be edited and the appropriate benchmark configuration to be executed should be commented/uncommented. By default, a single University dataset is configured and its configuration section is:

{noformat}# 1-university
[OWLIM_1]
class=owlim.OwlimWrapper
data=./univer1
database=jdbc:ignore
ontology=http://www.lehigh.edu/~zhp2/2004/0401/univ-bench.owl
{noformat}

One should use the {{\#}} symbol at the beginning of a line to comment it. The test should be launched via the {{lubm-benchmark}} script. Before either generation of datasets or execution of the test, OWLIM-Lite should be configured as discussed in the [installation section|OWLIM-Lite Installation] by editing {{lubm.ttl}}, however the file provided is suitable for running the benchmark with small datasets. To find out more about this benchmark, visit the [LUBM web site|http://swat.cse.lehigh.edu/projects/lubm/index.htm].

h2. Alternative to generating the LUBM datasets

It is also possible to generate the LUBM datasets 'on-the-fly' during the loading stage. In order to do this, edit the {{lubm.config}} file and use a data directory with the format 'GENERATE-n' where 'n' is the number of universities, e.g.

{noformat}# 1-university
[OWLIM_1]
class=owlim.OwlimWrapper
data=GENERATE-2
database=jdbc:ignore
ontology=http://www.lehigh.edu/~zhp2/2004/0401/univ-bench.owl
{noformat}

This approach saves the extra step (and disk space) for generating data files.