How to use it
Running the TPC-H is not so easy. Luckily we have developed a program which can help you to generate data, load and do the benchmark on your db. It works for the TPC-H and also for the SSB that will detail later on.
Download the program
wget [script_address]/database-benchmark.tar.gz tar xzvf database-benchmark.tar.gz cd database-benchmark
Data generation and loading
NB: Before executing database-benchmark, the db must already exist.
./database-benchmark -a load -m tpch
The “-a load” option is to use the program as a data loader and the “-m tpch” option is to work on TPC-H data model.
Perform a benchmark
./database-benchmark -m tpch
The benchmark performed by the program will allow you to select how many times you want each query to be executed. Then, it will display an execution summary for each request, and it stores it in a CSV file located in the working directory.
To use these results, you may have to do many benchmark on different solutions to make a choice. You also can compare your results to those published on the TPC website.
And that’s it!