3 ms·

> FF just seems to be getting slower It is a real pity, containers are awesome, one of the best new features in a browser for some time, but damn firefox is fr

by cheph 6y ago

> FF just seems to be getting slower 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.

bayindirh 6y ago
I think the magic word is seems. 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 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.
- 2Gkashmiri 6y ago
  where does pihole fit with this dnsmasq? i am running a pihole on a pi zero wh and that has just worked. can a pihole do this?
  - bayindirh 6y ago
    I don't know which DNS server PiHole uses or whether it's configured to cache normal DNS queries too. 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.
  - rand49an 6y ago
    Pihole is just a web app running over dnsmasq as far as I know.
- AmVess 6y ago
  It can be a lot slower depending upon the machine and OS it is installed on. For instance, running the Jetstream2 and Speedometer2 tests on https://browserbench.org 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.
  - bayindirh 6y ago
    Thanks for the quick benchmarks. 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.
rhn_mk1 6y ago
My Firefox setup blocks JavaScript by default. It has been peachy with NoScript and a bunch of other addons since forever. After Firefox switched to the new "Quantum" verion, my extensions got reshuffled and I'm using uMatrix to block scripts. 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.
- freedomben 6y ago
  I use both chrome and Firefox all the time to separate work and personal, plus dev testing (yes I have a lot of RAM :-D 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).
- paulryanrogers 6y ago
  My experience with Firefox quantum was the opposite. Ff became immediately faster and more useable. Experience seems to vary by hardware, sites, and add-ons.
- NikolaeVarius 6y ago
  I have 8 tabs open including a youtube tab, and FF is taking 0.5% CPU. Quantum was a "quantum" leap faster than what FF had become.
- cheph 6y ago
  I don't use it that often, but last I tried switching to it this was also one of my biggest problems, it was simply using too much resources when it was literally doing nothing, where chrome, doing much more, used much less CPU power. 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.

by cheph 6y ago

> FF just seems to be getting slower 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.

bayindirh 6y ago

I think the magic word is seems. 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 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.

2Gkashmiri 6y ago
where does pihole fit with this dnsmasq? i am running a pihole on a pi zero wh and that has just worked. can a pihole do this?
- bayindirh 6y ago
  I don't know which DNS server PiHole uses or whether it's configured to cache normal DNS queries too. 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.
- rand49an 6y ago
  Pihole is just a web app running over dnsmasq as far as I know.
AmVess 6y ago
It can be a lot slower depending upon the machine and OS it is installed on. For instance, running the Jetstream2 and Speedometer2 tests on https://browserbench.org 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.
- bayindirh 6y ago
  Thanks for the quick benchmarks. 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.

rhn_mk1 6y ago

My Firefox setup blocks JavaScript by default. It has been peachy with NoScript and a bunch of other addons since forever. After Firefox switched to the new "Quantum" verion, my extensions got reshuffled and I'm using uMatrix to block scripts. 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.

freedomben 6y ago
I use both chrome and Firefox all the time to separate work and personal, plus dev testing (yes I have a lot of RAM :-D 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).
paulryanrogers 6y ago
My experience with Firefox quantum was the opposite. Ff became immediately faster and more useable. Experience seems to vary by hardware, sites, and add-ons.
NikolaeVarius 6y ago
I have 8 tabs open including a youtube tab, and FF is taking 0.5% CPU. Quantum was a "quantum" leap faster than what FF had become.
cheph 6y ago
I don't use it that often, but last I tried switching to it this was also one of my biggest problems, it was simply using too much resources when it was literally doing nothing, where chrome, doing much more, used much less CPU power. 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.