Fastest ELPA

What’s the fastest ELPA? Let’s define it first! The ELPA runs the fastest? No, that doesn’t make any sense. The ELPA serves you the fastest? Maybe.

The definition for fastest ELPA in this article means it builds your packages the fastest! You might wonder why and how! Let me explain it to you!

Normal ELPA, like MELPA, builds packages from the recipes directory, so it takes O(n) to do so. The more packages one ELPA holds, take longer the time to get packages to update to the server. The general solution is to use a more powerful PC to reduce the build time.

Here comes JCS-ELPA, it uses the package github-elpa to host packages on the GitHub pages! But why is it the fastest LEPA? There are two factors:

  1. Microsoft owns GitHub so technically we serve our ELPA on Microsoft’s server
  2. We build packages using GitHub Actions with multiple jobs (if you are unfamiliar with GHA, think as multiple PCs helping you build your packages)

JCS-ELPA builds your packages with multiple jobs simultaneously! The formula:

New build time = Original Build time / jobs

If jobs is equal to 1, then the new build time equals to original build time (no build time improvement). For example, if original build time is 10 minutes, and we start 3 jobs, the new build time will be 2 minutes.

2 min = 10 min / 5 jobs

Of course, this is not going to happen in the real life. In reality, we will need to consume time on job scheduling, starting and closing time for the jobs, etc. But in theory, the more packages the ELPA holds, you will see the big differences between these two models.

👷 How does it work?

💡 Here we only explain how it works sine there aren’t enough packages to proof this ELPA is the fastest!

As of right now, there are only 136 packages hosted on JCS-ELPA. I have split these recipes into smaller sections. Let’s say 50 per section! The formula will be:

jobs = (total recipes / recipes per section) + 1


3 = (136 / 50) + 1

Now you have 3 workers to help you build packages onto our server! The original build time was 4 to 5 minutes, now it’s 1.5 minutes! What a big improvement! 😄


One of the coolest things is you can see what happens directly through the GHA’s log!