Stump's Performance Has Been Excellent!

[clseibold]

I just wanted to say that Stump is really excellent. It is one of the only self-hosting programs that has excellent performance, particularly in how it manages its database!

I've tried a lot of self-hosting things, like Komga, Emby, Jellyfin, Kavita, Universal Media Server, and they all usually have horrible database performance. But stump is very fast! I haven't checked into how you do it, but just wanted to say that that's the reason I love this program.

The UI is also really good. I do wish the comic book reader webapp thing was a bit better, but other than that, I love everything else.

[aaronleopold]

I'm happy you like Stump! If you have any metrics or performance readings you've taken, I'd love to see them. I'll also always suggest creating issues for any of the other apps you've mentioned having problems with (on their GitHub repositories) to sus out any potential config issues or to surface bugs that might improve the experience for everyone 🙂

I do wish the comic book reader webapp thing was a bit better, but other than that, I love everything else

There are some UI changes to the reader coming in the next release that have been sitting in nightly while my focus is here, but if you have any specific grievances feel free to capture them in an issue for review/discussion

[clseibold]

I don't have any performance readings aside from what I can visibly see, and the load time differences are visually apparent. Stump loads its pages in under 1 second, which is already better than all of the other media servers, which take up to more than 5x that time.

Emby and Jellyfin are already rewriting their database code (source: https://jellyfin.org/posts/efcore-refactoring). Since Jellyfin is a fork of Emby, it would not be surprising that Emby has (or had) the same database code that Jellyfin had before these EFCore refactorings.

As for Komga and Kavita, I have no clue what's wrong, but out of the box they were slow, so my assumption is just that the databases or SQL Queries are badly optimized. It's not really something a user can just pick out and write an issue on, especially if the problem is widespread in the database code. What I do know is that they are very slow, taking up to many seconds to load pages, let alone a long time to analyze the libraries. Stump, in contrast, has always been fast, and its pages load fast. And the library I tried stump with is the same as what I tried Kavita and Komga with.

From how it feels and looks, Stump just has a better optimized database, better concurrency/parallelization/threading, and/or a better optimized web frontend.

The performance issues with the others I'm sure are more apparent on slower machines, but the mistake devs might make is to then assume that speed is unrealistic on hardware from 2013. Stump proves this assumption incorrect. I also develop and run a search engine database with over 1 million rows in postgres on this same server, so I think that proves this assumption incorrect as well.

Anyways, thanks! I'll think about how to write the issue on the comic reader webapp when I have time.