The three most influential factors for Eclipse speed are:
Using the latest version of Eclipse (2020-06 as on 26 June 2020)
Note that David Balažic's comment (July 2014) contradicts that criteria which was working six years ago:
The "same" workspace in Indigo (3.7.2) SR2 loads in 4 seconds, in Kepler SR2 (4.3.2) in 7 seconds and in Luna (4.4.0) in 10 seconds. All are Java EE bundles. Newer versions have more bundled plugins, but still the trend is obvious. (by "same" workspace I mean: same (additionally installed) plugins used, same projects checked out from version control).
Launching it with the latest JDK (Java 14 at the time of writing, which does not prevent you to compile in your Eclipse project with any other JDK you want: 1.4.2, 1.5, 1.6 older...)
From version 4.8 (Photon) an up there was a steady speed gain after each version.
The main platform was optimized every release to load faster, enable more features for the dark theme and to add more features for newer Java versions for the Java development tools.
Especially with-in the last 3 versions the startup time was increased a lot. There should be a significant increase in start-up time with the newest version of Eclipse 2020-12.
In my experience it started a lot faster with each new version.
But: There are still plug-ins which do not follow the new way of using the Eclipse API and are therefore still slow to start.
Since the change to Java 11 as the minimum runtime version starting from Eclipse version 2020-09 at least the core system uses the newer features of the JVM. It is up to the providers of the other plug-ins to upgrade to newer APIs and to use the full power of modern CPUs (e.g. concurrent programming model).
Most of the answers you'll find around the internet will suggest you to either install the dependency to your local repository or specify a "system" scope in the pom and distribute the dependency with the source of your project. But both of these solutions are actually flawed.
Why you shouldn't apply the "Install to Local Repo" approach
When you install a dependency to your local repository it remains there. Your distribution artifact will do fine as long as it has access to this repository. The problem is in most cases this repository will reside on your local machine, so there'll be no way to resolve this dependency on any other machine. Clearly making your artifact depend on a specific machine is not a way to handle things. Otherwise this dependency will have to be locally installed on every machine working with that project which is not any better.
Why you shouldn't apply the "System Scope" approach
The jars you depend on with the "System Scope" approach neither get installed to any repository or attached to your target packages. That's why your distribution package won't have a way to resolve that dependency when used. That I believe was the reason why the use of system scope even got deprecated. Anyway you don't want to rely on a deprecated feature.
Since executing installation command for each lib is kinda annoying and definitely error prone, I've created a utility script which automatically installs all the jars from a lib folder to a project repository, while automatically resolving all metadata (groupId, artifactId and etc.) from names of files. The script also prints out the dependencies xml for you to copy-paste in your pom.
Include the dependencies in your target package
When you'll have your in-project repository created you'll have solved a problem of distributing the dependencies of the project with its source, but since then your project's target artifact will depend on non-published jars, so when you'll install it to a repository it will have unresolvable dependencies.
To beat this problem I suggest to include these dependencies in your target package. This you can do with either the Assembly Plugin or better with the OneJar Plugin. The official documentaion on OneJar is easy to grasp.