Benchmark S64 DA against PostgreSQL

Follow these instructions to run a TPC-H Scale Factor 1000 Benchmark comparing S64 DA and PostgreSQL.

This guide assumes that you have a database container instance already running. Otherwise please visit the S64 DA docker compose wiki to learn how to spin up S64 DA and PostgreSQL docker container instances the easy way.

Load the benchmark data

1. Download and unpack or clone the Swarm64 TPC toolkit.

2. Go to the unpacked or cloned directory.

3. Install the required Python packages with pip3 install -r requirements.txt.

4. Install the PostgreSQL 11 client.

5. Change directories: cd schemas/tpch.

6. Run the TPC-H data loader to prepare the S64 DA database:

   ./loader.sh --schema=sdb_all_tables --scale-factor=1000 --dbname=tpch_sf1000_s64da

   This loads a TPC-H dataset with 1TB worth of data.

7. After the data-loading process is complete for S64 DA, load the PostgreSQL database with the same dataset:

   ./loader.sh --schema=psql_native --scale-factor=1000 --dbname=tpch_sf1000_psql

   You have successfully loaded the database schemas. Depending on the scale factor and hardware that you used, the loading operation can take up to several hours to complete.

Execute the benchmark

1. Change directories to the root directory of swarm64-tpc-toolkit.

2. Run the benchmark on S64 DA with a timeout of 15 minutes:

   ./swarm64_run_tpc_benchmark --dsn postgresql://postgres@localhost/tpch_sf1000_s64da --benchmark tpch --timeout 15min

3. After completion, run the same benchmark but on PostgreSQL:

   ./swarm64_run_tpc_benchmark --dsn postgresql://postgres@localhost/tpch_sf1000_psql --benchmark tpch --timeout 15min

If you encounter any issues, please send an email to [email protected]