Slow search queries when running pgvector with external database

[gabrieloqm]

Hello everyone,
I am running ann-benchmarks using pgvector algorithm/extension on different datasets provided by this repository and with different parameters for the hnsw index. At first, I have tested it on my local machine, running the run.py script. But then, i have tested it on a external database instead of initializing and creating a postgreSQL database which runs on the docker container initialized by the script mentioned above.
I have enhanced some of the postgreSQL external database configurations, as you can see below (the comparison between the database created by the ann-benchmark script and my external database):

What I am facing is that the qps metric on my external database is way worse(100~ times slower) than the one I get when running the database locally, even though the external database configuration is more robust.
Any tips on what could it be, or rather, what configuration could i change to improve queries speed?

[SunnyBo0302]

It might be a network-related issue.

Could you try running ann-benchmark on the server that hosts the external database directly, to see if the problem persists?