this post was submitted on 09 Sep 2022
18 points (100.0% liked)

General Programming Discussion

7814 readers
1 users here now

A general programming discussion community.

Rules:

  1. Be civil.
  2. Please start discussions that spark conversation

Other communities

Systems

Functional Programming

Also related

founded 5 years ago
MODERATORS
 
top 6 comments
sorted by: hot top controversial new old
[–] ttmrichter@lemmy.ml 5 points 2 years ago

Every programmer?

Methinks somebody doesn't know how broad and deep the field of programming is...

[–] Helix@feddit.de 2 points 2 years ago (1 children)

And why? What do I gain from knowing what latency something produces?

That's just a rough estimate and not a real world number I get from a deployment of my software on my own hardware or some cloud service. What do I gain from knowing that the RTT from CA to NL is 150ms compared to a memory reference of 100ns?

Everyone knows local data is faster than remote data and SSDs are faster than HDDs and that the closer you get to the CPU, the less latency you can expect.

[–] Ephera@lemmy.ml 3 points 2 years ago (1 children)

I mean, sometimes knowing 'normal' timing values can help you debug stuff.

For example, we recently were debugging why we sometimes didn't get data in time from a network sensor. And we probably wouldn't have thought much about the logs saying it took 300 ms to ping that sensor, if we didn't know 50 ms is enough to ping many webpages.

But yeah, unless you're a performance tester, I cannot imagine why you would need to know all of these values, rather than just their rough relation to each other.

[–] Helix@feddit.de 1 points 2 years ago (1 children)

but can't you just ping some other thing you know works properly adjacent to that sensor? So for example, if you have two servers, one behaving normally, the other weirdly, just compare them?

[–] Ephera@lemmy.ml 3 points 2 years ago

You could do that, if you're actively checking ping roundtrip times. It was in a local network and generally, the connection was working, so we didn't think the ping or roundtrip times would be relevant.

Our software was just routinely logging ping+RTT and when scrolling through the logs, we noticed more or less by chance that the RTT is 300 ms, which is absurdly high for that context.

And well, this is just one example. If you're doing time-sensitive stuff, it's useful to know the timings of what you're dealing with. It's not usually essential, and you especially don't need to know it for every possible context. But it can make your life easier.

[–] dwzero@lemmy.ml 2 points 2 years ago

Most of this doesnt seem like something i'll use, but i appreciate it all the same.