ants_a's comments

On: Do low-level optimizations matter? Faster quicksor...

[ants_a]

No, and it feels unlikely that they will either.

On: Lessons Learned from Scaling to Multi-Terabyte Dat...

[ants_a]

For RDBMS we typically consider millions to be tiny to small. A billion is somewhere on the boundary of medium to big.

On: How many EV charging stations does the US need to ...

[ants_a]

Traffic on I-10 at Arizona state line is 31k vehicles per day, with peak hour being 4.3k vehicles per hour. Assuming 350Wh/km, 120kW average charging rate and 90% utilization 3.2 charger stalls per mile per 1k vehicles/hour are needed. Or putting it differently, a 100 charger site with 12MW of power every 5 miles.

[jmpman → ants_a]

Is that peak for a normal day, or for Thanksgiving weekend?

On: pg_timeseries: Open-source time-series extension f...

[ants_a]

    SELECT date_bin('5 minutes', log_time, '2000-01-01') log_time,
           AVG(value) avg_value
    FROM logs GROUP BY 1

On: Tesla's wafer-sized Dojo processor is in productio...

[ants_a]

A comparable product from Cerebras is rumored to cost around $2-3M.

On: PCIe 7.0 Draft 0.5 Spec: 512 GB/s over PCIe x16 On...

[ants_a]

That's not the motherboards fault, the PCIe controller is on the CPU and needs to have corresponding I/O pinout on the socket. And even from the CPU manufacturer side, it's not artificial market segmentation. The I/O die alone on an EPYC/ThreadRipper Pro is more than double the die area of the CPU, GPU and I/O chips in a Ryzen 7 7800, and it's on a single chip which makes yields way worse. The socket and package need 1700 connections in one case and 6000 on the other. The high I/O CPU is way more expensive to manufacture, and on the lower end the prices start at ~1k for 512 GB/s of full duplex PCIe bandwidth and 12 memory channels.

On: 6.2 GHz Intel Core I9-14900KS Review

[ants_a]

For well scaling workloads optimizing for efficiency is optimizing for performance. For serial workloads even one P-core is enough. There aren't too many workloads in the middle that scale but only to a few cores.

On: SSDs have become fast, except in the cloud

[ants_a]

Networking doesn't have to have high latency. You can buy network hardware that is able to provide sub-microsecond latency. Physical distance still matters, but 10% of typical NVMe latency gets you through a kilometer of fiber.

On: Show HN: Easily Visualize Your SQLAlchemy Data Mod...

[ants_a]

Nice, I hacked together something similar many moons ago, it looks like some fine folks have kept on maintaining and improving it: https://github.com/fschulze/sqlalchemy_schemadisplay

Useful to have in a CICD pipeline so you always have an up-to-date schema diagram available.

[eigenvalue → ants_a]

Cool, I'm surprised I couldn't find this when I was looking for something like this originally. Probably because I was using different terminology-- "data model" instead of "schema," and "visualize" instead of "display".

On: Introducing PGMQ: Simple Message Queues Built on P...

[ants_a]

Is there a reason why this is implemented in Rust instead of plpgsql? At first glance I didn't see it using any features that would need a loadable extension. Plain SQL based solution would make using it on managed database services much easier.

[chuckhend → ants_a]

It probably could be implemented as pure plpgsql for some of its features (basic send/receive). The long-term plan is to add additional message type support (protobuf, msgpack, etc), embed a metrics exporter (prometheus?), push notifications, and. for those we want to pull in Rust modules from open source. I bet the project could be refactored to expose the parts you're looking for as plpgsql though.

Distribution of Postgres extensions is a challenging problem that we're working on over in https://github.com/tembo-io/trunk // https://pgt.dev/. One of the goals there is to make it easier to install extensions.