In another thread in this forum, I mentioned my previous browser benchmarks, and I said there that I thought I’d try them again. I did, and here are the results!
I changed the methodology a bit. Previously, if I remember, I ran each test with the browser configured as I would use it, with the full complement of addons. I used the same addons on each one where possible, and the closest ones I could find if the same one was not available for other platforms (webextensions or Chrome). That’s pretty sloppy, really, as there’s no way to know how much of the browser differences is from the browsers themselves and how much is the addons. I think I may have tried with and without addons when I first ran this “shootout” right after Firefox Quantum landed, but it’s been a while, and memory fades.
Right now, I ran with new profiles on each browser. In Waterfox and Firefox, I force-enabled hardware acceleration, which is off by default on Linux. Firefox defaults to 9 content processes, so I used that, and in Waterfox, I set it to use the maximum, which is 7. Chromium-based browsers use one process for each tab, so there is no actual maximum, as far as I know.
The test was SpeeDOMeter 2.0, from browserbench.org. This is the same test that Mozilla used to claim (accurately) that Firefox was now twice as fast as it had been some time ago. I repeated their test back then and found the same results they claimed. Because this is the “official” test for the Mozilla claims, I use it for these tests (whose main purpose is to see if the speed benefits claimed for Quantum have materialized).
In this case, I used my Dell G3-3579 (15.6 inch display). CPU is an Intel i7-8750H, with 16 GB RAM, with the nVidia GPU enabled in Linux, driver 440.59 (proprietary). OS is KDE Neon Linux 5.18.2, kernel 5.5.6 mainline, Mesa 20 (Kisak PPA). The results in Windows should be the same as these with the same hardware.
Each browser was maximized and with the zoom set to 110%.
First up was Waterfox Classic, 2020.02.1. It scored a respectable 65.6, considerably higher than it does with my addons installed (about 40 in that case).
Next was Firefox 73.0.1. It scored an impressive 77.09, giving it a significant lead over Waterfox. That’s an 18% gain, the highest I have ever recorded in these tests.
Finally, I ran the test on Chromium 81.0.4021.2 (Developer Build). It scored an even more impressive 86.3, representing a 12% gain over Firefox.
Firefox has opened a significant lead over Waterfox in performance. There’s no way to know whether any of the changes Mozilla made to achieve this depend on not having the classic addon API(s), unless we ask a Mozilla dev, perhaps.
Once again, though, Chromium wins on performance. In terms of my preferences, though, I’d pick Waterfox first, Firefox second, and Chromium last, exactly opposite of the rankings in performance. Waterfox is the slowest, but it’s fast enough. Once a browser is fast enough, features make the most difference to me, and even if I am certain that the browser is de-Googled, if necessary (as it would be with any Chromium variant), other factors, like the UI configurability, mean more than the performance. Others may disagree, of course!
Dell XPS 13/9310, i5-1135G7/16GB, KDE Neon 6.2
XPG Xenia 15, i7-9750H/32GB & GTX1660ti, Kubuntu 24.04
Acer Swift Go 14, i5-1335U/16GB, Kubuntu 24.04 (and Win 11)
