Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Shell Python
tree: a6e56defe2

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.


Berlin SPARQL Benchmark Automated

Berlin SPARQL Benchmark (BSBM) is a benchmark for comparing performances of an 
RDF store across different architectures or for comparing different RDF stores.

The aim of this project is to make as easy as possible for people to run the 
Berlin SPARQL Benchmark with Joseki [1], Fuseki [2] and (this has not been 
implemented yet) a local TDB [3]. Other RDF stores might be added in future.

You can simply run the bash script typing:


You need bash, Java, Ant, Maven, SVN, wget, etc., and I am not going to explain 
how to install/configure those. The script downloads all the necessary software 
pieces, it sets them up, it uses the BSBM to generate a test dataset and it 
runs the benchmark for you against Joseki/TDB and Fuseki/TDB.
Once finished, you can find the results in the /tmp/bsbm/results/ directory.

You can take a look at the script and, once you have run it a couple of 
times, I suggest you change the parameters at the beginning a start doing
some serious benchmarking, in particular:


You might want to use tdbloader2 instead of tdbloader:


I am not a bash "guru", I warned you! You can insult me or send me suggestion
how to improve the script. Suggestions are more welcome than insults. ;-)

See also:



Kudos to Chris Bizer and Andreas Schultz for having written and shared the 
Berlin SPARQL Benchmark (BSBM) which although not perfect, as any benchmark, 
can brings a degree of objectivity and replicability in the world of RDF store 
performances. Kudos to Andy Seaborne too for TDB and Fuseki and the continuous 
effort towards improving them as well as their performances.     

Have fun with the BSBM and TDB/Fuseki!

                                                              -- Paolo Castagna

Something went wrong with that request. Please try again.