TPC-B-like test
The scaling diagram has the same basic curve on the write-heavy test, but with fewer transactions occurring:
If you look carefully at the left side of this, you can see one of the issues warned about above and in the pgbench documentation: results at a very low database scale are lower than when the database is a bit larger. This is due to lock contention for the small branches table in particular, when running at higher client counts with a tiny scale.
The above results are averaged across all the client counts used. The number of clients also impacts results quite a bit:
Corresponding, the client results are also averaged out over all the scaling factors. The client scaling for the read-only test looks similar, just with smaller total values.
To get a really clear picture of the system, you should also take a look at the web report produced by pgbench-tools. The number of clients and database scale that produced the greatest single TPS value is usually an interesting...
