abp vs µBlock: Acid3 test

Giải Trí

Adblock Plus vs. µBlock.

Test site:

This is a *very simple* web page, with no ads, no trackers, and no analytics embedded.

The purpose of the test is to show the delay caused by ABP’s unconditional injection of over 13 thousands CSS rules (if using only EasyList + EasyPrivacy). This test shows ABP’s overhead vs. µBlock’s overhead with regard to element hiding filters. (µBlock’s overhead with regard to net request filters is also significantly lower, though the test here is not a good one to demonstrate this given the small size of the page).

First three clips are three distinct tests. The fourth clip is the first clip in slow motion. The icon in the tab was used to synchronize the videos.

Whereas ABP injected 13,900+ CSS rules in that very simple web page and each of its embedded frames, µBlock injected not a single one (µBlock surveys the web page at load time to find out which element hiding filters are needed, if any).

So as a result, µBlock’s memory footprint is reduced, as well as CPU footprint, which is why ABP lags in this benchmark at page load time. Keep in mind that lower CPU footprint also translates into less power usage, which is something to consider for devices running on batteries. (Example: The test here was performed on a performant CPU (i5-3570K CPU @ 3.4GHz x 4).

Even with powerful devices, the CPU will typically runs cooler with µBlock when I submit both ABP and µBlock to my reference benchmark:

If you happen to use more lists than the ones used for this tests (EasyList + EasyPrivacy), expect ABP to cause more lags + resource consumption, while µBlock will just behave the same (because no rules are injected if not needed).

Even for pages with ads and/or trackers and/or analytics, µBlock will inject a tiny fraction of what ABP injects on each page and frames in a page.

Nguồn: https://coca-colafoundation-ind.org

Xem thêm bài viết khác: https://coca-colafoundation-ind.org/giai-tri

Leave a Reply

Your email address will not be published. Required fields are marked *