Surprising Result in Browser Speed Experiment

[PaulPA]

I thought I would share this little experiment with the WW community.

I have been a loyal Firefox user for many years. Recently, the work I am doing on upgrading to a new version of my CMS has led me to question my loyalty. The issue is speed. Specifically the length of time it takes to save an item was frustrating me. I am on a dedicated server that is fairly quick and so is my network speed, so I had to look at whether there was something going on with the browser. So I conducted an experiment. Now this is highly unscientific, but results may still raise some eyebrows.

The experiment involved simply measuring the time it takes to save an opened item (in this case a menu module). No new data was entered. The item was simply opened from the module main page and then a button was clicked that saved and closed the item. A timer button (on a smartphone) was pushed to start when the Save & Close button was clicked and was pushed to stop when the module main screen reappeared. This was done 5 times for three different browsers: Firefox (14), Chrome (21) and Internet Explorer (9). Additionally, the tests were alternated for each round so that each browser would be tested in each round before moving to the next round.

Results were measured in seconds and here are the averages:

Firefox 13.8 sec.
Chrome 7.3 sec.
IE 12.3 sec.


The results were very surprising to me. And because I have many items to change, the potential time I may be able to save could prove significant.

I am wondering if anyone else has experienced similar results.

[dstiles]

Possibly slowed by too many or wrong type of add-ons? That is a known feature of FF.

Also, if something like (eg) "check URL for virus" is enabled in FF then that could be very time-consuming.

[Panthro]

I've never done any tests, but I stopped using FF a while back bc it was too slow, kept eating up RAM and crashing.

[Frank_Rizzo]

Add-ons are a big cause of slowdowns / memory hogging with Firefox.

Mine is loaded with ad blockers, noscript, ghostery, https-everywher, cookie privacy etc. But in this instance I would go 100% over Chrome even if it meant production was 50% slower.

Don't forget that you can optimize Firefox for better performance. Look at add-ons fasterfox / fasterfox lite.

Oh. And I doubt Chrome would allow add-ons such as OptimzeGoogle? :-)

---

I like the way FF 15 (I think this is the version that started it) allows you to enable / disable add-ons on the fly without restarting. This is essential if you want to check your adsense reports.

[PaulPA]

Just for fun, I disabled all FF add-ons and restarted the computer. Ran the test again 5 times for FF and Chrome with similar results:

Firefox 13.6 sec.
Chrome 7.3 sec.

[Frank_Rizzo]

How about uninstalling them rather than disabling?

And then trying with Fasterfox / Fasterfox Lite? (although that is more of a network optimiser)

[robzilla]

A stopwatch (or smartphone equivalent) is a poor measuring device. Both Firefox (with Firebug add-on) and Chrome (built-in Developer Tools) allow you to see what resources are loaded, and how long this takes. Compare those results, as well as the headers sent and received, and you might find the culprit.

If I recall collectly, Chrome is slightly better than Firefox in terms of Javascript performance. If your CMS is script-heavy, that might explain the difference.

It does not have to be a rendering issue, though. It could be that your CMS treats browsers differently.

[PaulPA]

If I recall correctly, Chrome is slightly better than Firefox in terms of Javascript performance. If your CMS is script-heavy, that might explain the difference.

You may have something here. I checked and there seems to be an issue with how FF is handling the javascript. However, I would argue that Chrome is not "slightly" better. At half the speed it is significantly better considering how many updates I need to make and how much time it will be saving.

[Seb7]

A stopwatch is a poor measuring device

yes, but still significant, i much doubt he was 5 seconds off que.

It depends what you are doing with these browsers. Each of them has something that performs better than the others. and they all are getting better - some more than others.

I currently prefer chrome (on pc) and think its generally the fastest. The mac may not be as good. But sometimes I switch to ff and do find somethings are plainly faster. I think IE developers would be embarrassed to see the fanfare that microsoft makes of IEs performance.

[motorhaven]

I take browser benchmarks with a grain of salt these days. I use Opera, Firefox, Chrome and IE. I've seen all of them beat the others depending on what's being loaded. The most important measurement of speed I have as a web admin is that they all load quickly on my sites. If any of them do not, I track the reason why (even IE has build-in developer tools) and try to address it.

What I use elsewhere is most often Opera, for usability reasons.

[ergophobe]

I have one computer where Firefox is so slow, I have to use Chrome unless it ends up too slow and I have to use IE or Opera. I have another where Chrome seems slightly slower than Firefox, but in general, all the browsers are about the same.

In other words, I think there are a lot of factors that affect this and it's hard to pin it on any one thing (unless you find an offending plugin or something).

[Brett_Tabke]

Load and render time of a single page isn't the real metric of speed. I think you have to factor in the entire surfing experience.

If you know how to use Opera well, I believe it is significantly faster than any browser mentioned in this thread.

Some things I do:
- Rarely open a page with a single 'click'. When using Google, just SHIFT-Ctrl-Click rapid fire all the results into new background tabs. Read the first one while the rest are loading.
- Use the 'single key' inputs. (z=back, x=forward).
- Use mouse gestures. Navigation doesn't require keyboard 90% of time.
- Use bookmark branches and open 20-100 tabs with one click. This is great for your "morning runs". Never forget to check a site again!
- Privacy-Security: turn on/off referral strings/cookies/plugins by pressing f12.
- Bandwidth stealing ads: turn off ad loads by right clicking on them.

[ergophobe]

I've been defaulting to Opera more and more these days because, strangely, it seems to be the most stable for me these days. I say strangely because of course most sites don't even test for Opera, but I often have trouble with Google properties in both Firefox and Chrome. On my old, slow desktop Opera and IE are way faster than FF and Chrome. I expect that's because of some sort of conflict or problem with the JS engine who knows, but even with all the plugins off, Firefox bogs down all the time (on the laptop, all browsers are fast and stable) and Chrome tends to be unusable even on Google properties.

I haven't quite gone to OPera as my default browser, but I do think the clunky feel I used to dislike has given way to a far less clunky feel than Firefox, currently the clunkiest feeling browser on the market.

Firefox is a great tool, though - Firebug is still awesome, HTTP Live Headers, Tamper Data, Web Developer are all handy.