PostgresML is 8-40x faster than Python HTTP microservices - PostgresML

[giscus[bot]]

PostgresML is 8-40x faster than Python HTTP microservices - PostgresML

PostgresML v2.0 outperforms Python HTTP microservices by a factor of 8 in throughput, latency, and memory utilization.

https://postgresml.org/blog/postgresml-is-8x-faster-than-python-http-microservices/

[jfmatth]

I'm not an ML guy, but love to hear the Postgres ecosystem is doing great here!

[arvsr1988]

This is not a fair comparison. You need to serve the postgres query over a http server just like the python application.

[montanalow]

I think the point the post illustrates beyond the straight benchmarks, is that you actually don't need an http service to access ML with PostgresML. Connectivity comes from PostgreSQL instead. It's a completely different stack with a completely different performance profile.