5

I've seen that some people have compiled themselves so-called "optimized" versions of Firefox, from the official source code.

One example of such an optimized build is the "Blazing Fast", from Binary Turf. They have versions 3.0.11 and 3.5.2 of Firefox optimized for CPUs supporting SSE and SSE2.

If you have a supported CPU (I guess nearly everyone can run the SSE2 version, as it's for Pentium 4 and above), is it worth installing it over the standard official Firefox release.

What are the advantages? Better speed loading and scrolling the pages? Javascript speedup? Memory usage (I doubt it)? Anything else?

1
  • 1
    I highly doubt you're gonna see any major improvements.
    – alex
    Sep 8, 2009 at 13:41

5 Answers 5

7

BinaryTurf posted some numbers here.

Firefox, default on SunSpider:

Total:                  5306.4ms +/- 0.9%

Firefox, SSE2 optimized on SunSpider:

Total:                  5295.6ms +/- 1.7%

A net 0.2% speed improvement IF the numbers were spot on. They don't make a strong case for themselves.

There are SSE2 optimizations targeted at highly specialized desktop applications like Lightwave or specific Photoshop rendering algorithms, but Firefox has little in common with those. The bulk of my personal usage is spent fetching text via HTTP, or waiting for my internet connection to load pages. I can't think of a worse situation for an SSE2 optimization, and the results above seem to bear that out.

1
  • 1
    certainly nothing to write home about :)
    – Molly7244
    Sep 8, 2009 at 15:07
5

On linux builds it almost definitely DOES make a difference because many popular distribution packages of firefox are built without Profile Guided Optimization.

Check out Swiftweasel, a tricked out firefox/mozilla build for linux.

2
  • I'm running Windows and thus more interested in optimized builds for Windows but it will surely be interesting for the Linux folks.
    – Snark
    Sep 8, 2009 at 14:26
  • I'd like one as I suspect my VIA C7m is weird enough to make a default set of optimizations less than ideal. All I need is the right little snippet to benchmark until I find the way that works best (too long to build the whole thing) Sep 8, 2009 at 21:36
5

In a nutshell: no.

Never believe any of the myths people spread on the net about browsers optimization techniques. Keep in mind that if there were a way to improve Firefox performance noticeably, the Mozilla folks would have already implemented it in the code.

1
  • Compiler optimization doesn't work that way. A binary can be optimized for specific target system (mostly CPU) and yield significant boost. But a central official distribution generally doesn't use them because they need to distribute universal binary that will work on any system. Feb 13, 2021 at 3:37
1

I never test those version, only heard about them and I also doubt that you'll see major improvement.

Firefox 3.5.x branch already brings major performance boost.

Also before using those "modify" version, be aware of what has been done in it to "boost" them to be sure it can't be harmful such as integrated spyware...

Many OpenSource software are modified and this is not always for the best, just take the exemple of OpenOffice where many modified version are online and some are modified just to make users paid to use it, that is just BS because OOo is a free sotfware and some make people pay for it..

So I believe that some crooks may do the same with firefox...

1

If you can understand what this does inside mozconfig:

ac_add_options --enable-optimize="-Ox -GL -GY -GAFs -arch:SSE -fp:fast"

you can certainly understand what's optimized. Otherwise you can only argue.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.