A little more than a year after releasing the first version of Firefox,

Mozilla.org has finally released Firefox 1.5. This new version is supposed to

add all sorts of new features including Live Bookmarks which merge RSS and

bookmarks, improved error pages, improved patch mechanisms, as well as more

security and faster load times. Naturally, with all these new features, you’d expect it to be less efficient.

I started by loading Firefox 1.5 on my Linux

workstation. I had a few conflicts with packages loaded by prior versions, but

it was nothing that was Firefox’s fault. Once it was finally up and installed I

checked to see how big of a footprint Firefox 1.5 had over Firefox 1.07. Here’s

what I found:

  • Firefox 1.07 showed RAM usage was 38.7MB, with 78.6MB of Swap.
  • Firefox 1.5 showed RAM usage was 40.7MB, with 78.4MB of Swap.

As you can see, Firefox 1.5 is a little bigger on my

test machine than Firefox 1.07. But this was on a Linux machine and your mileage

may vary anyway. The added bulk hasn’t seemed to have negatively impacted

performance though. Firefox 1.5 seems to load a little quicker and load pages

quicker than 1.07 as well.

On my Windows XP workstation, Firefox 1.5 consumes 20.5 MB of RAM

before visiting the first page. The real proof in the pudding however

is how it performs as you use it. Visiting the TechProGuild homepage

immediatlely causes Firefox’s usage to climb to 24.8 MB on Windows and

42.3 MB on Linux. Load a couple of articles and read them, and the

memory usage continues to spiral. Firefox 1.5 occasionally releases the

memory but not as fast as it should. There have been occasions on my

Linux box where Firefox has grabbed over 150MB of RAM and wont let go

even though there’s only one window and page active. Normally

however that’s after I’ve had several windows and dozens of tabs

running.

It’s interesting that Firefox uses less resources under Windows than

it does under Linux. I checked Erik Eckel’s Mac and Firefox consumes

more memory there as well. I’m sure this has a lot to say about

the robustness of the development tools.