When dealing with joins, table size can become critical regarding speed. Postgresql provides a set of variables that can be modified (depending on system resources) to help the performance.

The goal of this talk is to provide some initial benchmarks given the default configuration over a set of joins (each using an increasing number of rows), and contrasting those results with a series of tuned parameters (using the same queries on the same hardware).

This will help provide some information as to what impact can be hoped for, as well as provide some information as to when a table becomes 'too big'.