Skip to content
This repository has been archived by the owner on Sep 9, 2022. It is now read-only.

Myth: µBlock consumes over 80MB

Raymond Hill edited this page Aug 13, 2014 · 19 revisions

I've seen a couple of instances of people claiming µBlock is not as memory efficient as claimed. Examples:

When µBlock launches, it loads all selected filter lists, parse the content, eliminate duplicates, then instantiates the filters using efficient internal representation. This parsing of the filter lists requires a good amount of temporary memory.

So if you look at the task manager right after µBlock has loaded and parsed the filter lists, you will still see µBlock's memory footprint as a result of loading all the filter lists. Still, at this point all this temporary memory has been relinquished to the browser, but the browser hasn't yet collected the freed memory to make it available for reuse.

If the browser is idle enough, before one minute has elapsed, the browser should be able to garbage collect the temporary memory which was freed by µBlock after it finished loading and parsing the filter lists:

µBlock's memory footprint

The top image shows the memory footprint of µBlock right after launch (Chrome 64-bit) (Expect similar memory footprint each time the filter lists have to be reloaded). The image at the bottom shows the memory footprint of µBlock before one minute has elapsed while the browser is idle.

Note that µBlock's baseline memory footprint won't change that much afterwards. It will likely settles a few MB above the memory footprint reached after garbage collection has occurred, whenever the garbage collector is permitted to do its job.

Clone this wiki locally