It is a real pity, containers are awesome, one of the best new features in a browser for some time, but damn firefox is frustratingly slow.
It is a real pity, containers are awesome, one of the best new features in a browser for some time, but damn firefox is frustratingly slow.
My experience with Firefox got much faster after I installed a local DNSMasq server.
Things got noticeably faster after that point.
For details, see: https://news.ycombinator.com/item?id=21877330
This comment was when I first installed DNSMasq on a very low powered OrangePi Zero (512MB) on home network. I worked a lot of kinks and this situation holds truer when compared to that time.
For instance, running the Jetstream2 and Speedometer2 tests on https://browserbench.org on Mac OS M1 CPU and Windows 10 Ryzen 3700:
Mac OS M1: Jetstream2 Firefox score of 115. Safari score of 227. Speedometer 2 Firefox score of 203. Safari score of 218.
Windows 10 Ryzen 3700: Jetstream2 Firefox score of 92. MS Edge score of 145. Speedometer 2 Firefox score of 95. MS Edge score of 142.
These are bare metal installs running off the same connection method to the same network. I understand what you are saying about DNSMasq, but I wanted to test out the "seems" portion of his comment. If Firefox seems slower on any particular machine, that is because it is.
Since I mainly use Linux (80%) and Mac (20%), I sometimes think that everyone is using the same OS with me.
You're right, I didn't think of OS dependent performance difference.
The aim here is to minimize the DNS query turnaround time and reduce the latencies while opening a connection.
When the network is working smoother, Firefox's performance goes up in a disproportional fashion. I didn't notice the same amount of improvement in other network bound stuff but, Firefox went from "meh" to "youjustdidwhat!?" after DNSMasq installation and modem change.
Same thing also happened when I changed my modem 7-8 years ago after my venerable Alcatel Speedtouch died.
So, in retrospect, modem chipset really plays a central role in your home network performance.
However, even with scripts blocked, Firefox eats up a whole lot of CPU power, even on static, unchanging pages. What was <1% is now 5% CPU.
Given the Quantum promise of a speedup, I'm rather underwelmed.
I don't like chrome, firefox is feature wise better IMO, but I am not patient enough for it.
Maybe it is just my setup, and the plural of anecdote is not data, so my experience may be an outlier, but I wish it was not my experience.
It seems to be that Firefox uses more cpu and better on memory, whereas chrome is the opposite (although by by means is chrome "light" on cpu).
Quantum was a "quantum" leap faster than what FF had become.