Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#1

Post by gaocegege » Mon, Dec 08, 2025, 5:30 PM UTC

Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM
blog.vectorchord.ai

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#2

Post by duckbot3000 » Tue, Dec 09, 2025, 3:38 AM UTC

Kinda makes you wonder why you need cloud for anything besides remote encrypted backups if you can run all that on 12GB

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#3

Post by setr » Tue, Dec 09, 2025, 6:23 AM UTC

duckbot3000 wrote: post #2

Kinda makes you wonder why you need cloud for anything besides remote encrypted backups if you can run all that on 12GB

Because getting any hardware out of infra-team on premise is utterly miserable, across the board.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#4

Post by nwellinghoff » Tue, Dec 09, 2025, 7:46 AM UTC

Too bad aws does not support any of these other vector extensions in managed rds.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#5

Post by riku_iki » Tue, Dec 09, 2025, 9:11 PM UTC

duckbot3000 wrote: post #2

Kinda makes you wonder why you need cloud for anything besides remote encrypted backups if you can run all that on 12GB

what about failover story if server dies? PG failover setup is complicated, and cloud infra handles this for you.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#6

Post by ayende » Sat, Dec 13, 2025, 5:22 PM UTC

That suffer from a serious issue

You must have the data upfront, you cannot build this in an incremental fashion

There is also bo mention on how this would handle updates, and from the description, even if updates are possible, this will degrade over time, requiring new indexing batch

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#7

Post by positron26 » Sat, Dec 13, 2025, 5:47 PM UTC

riku_iki wrote: post #5

duckbot3000 wrote: post #2

Kinda makes you wonder why you need cloud for anything besides remote encrypted backups if you can run all that on 12GB

what about failover story if server dies? PG failover setup is complicated, and cloud infra handles this for you.

Do we mean managed or PG on K8s like CNPG? In all cases, I use the infra to simplify things like having disk redundancy and failover nodes, not because 12GB is interesting.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#8

Post by lelanthran » Sat, Dec 13, 2025, 5:50 PM UTC

setr wrote: post #3

duckbot3000 wrote: post #2

Kinda makes you wonder why you need cloud for anything besides remote encrypted backups if you can run all that on 12GB

Because getting any hardware out of infra-team on premise is utterly miserable, across the board.

That's not the only alternative.

Rent your VPS and add in extra volumes for like $10 per 100GB.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#9

Post by riku_iki » Sat, Dec 13, 2025, 5:51 PM UTC

positron26 wrote: post #7

riku_iki wrote: post #5

Earlier quoted context omitted.

what about failover story if server dies? PG failover setup is complicated, and cloud infra handles this for you.

Do we mean managed or PG on K8s like CNPG? In all cases, I use the infra to simplify things like having disk redundancy and failover nodes, not because 12GB is interesting.

Primary managed PG, since you still need setup/maintenance/monitoring on your K8S own solution.

Re: Indexing 100M vectors in 20 minutes on PostgreSQL with 12GB RAM

#10

Post by esafak » Sat, Dec 13, 2025, 6:08 PM UTC

How does it compare with paradedb and lancedb?