Testing Opera and Firefox browsers + Firefox Builder Optimization utility

Introduction

I decided to conduct another testing of Opera and Firefox browsers for myself. I myself am a fan of Opera products. But don’t mind touching other convenient browsers, which Firefox is for me (the rest aren’t appealing to me, it may just not be enough). It is also sometimes “to touch” unfortunately is forced, as, for example, with the service of Google Wave . In addition, browser usage statistics in runet confront these competitors with their brows (approximately 25% for FF 3 and Opera 9-10, I don’t take IE into account - you understand that there is no sense in it, although the share is 6, 7 and 8 its versions are large enough).




')
I don’t really believe other statisticians . Of course, they may have it there in Germany, but it’s hard for me personally to believe in Firefox’s total superiority over its long-time rival Oper ...

So let's start with ...

Data

Browsers :
For the test, I took stable and test versions of browsers, or rather Opera 10.10 build 1893 (I use it daily), Opera 10.50 Internal build 3199 (news here ), Firefox 3.5.7 , Firefox 3.6 pre (Namoroka) , Firefox 3.7 a1 pre (Minefield) .

OS, iron :

operating system	Microsoft Windows 7 Home Premium
CPU type	Intel Core 2 Duo E6420, 2133 MHz
Motherboard	Asus p5b
Motherboard chipset	Intel Broadwater P965
System memory	2048 MB (DDR2-800 DDR2 SDRAM)
Video adapter	NVIDIA GeForce 8600 GT (256 MB)

Internet and processes :
Internet channel symmetrical 10 Mbit. All the programs are completed, the antivirus is deactivated. Browsers opened one by one, tested and closed.

Tests :
Conducted 2 tests, they probably all know them: the Peacekeeper integrated test - The Browser Benchmark from Futuremark and the SunSpider JavaScript Benchmark script processing from the WebKit project.

Additions :
I also used Firefox Builder Optimization (hereinafter referred to as FBO). I tested it with and without it (everything is displayed in the table below). By the way, I could not find a single mention of him on Habré.

results

	Peacekeeper without fbo	Peacekeeper with fbo	SunSpider without fbo	SunSpider with fbo
	points (more is better)	points (more is better)	ms (less is better)	ms (less is better)
Opera 10.10 build 1893	1817	-	3741.4	-
Opera 10.50 Internal build 3199	2877	-	559.6	-
Firefox 3.5.7	1797	1764	1395.4	1368.6
Firefox 3.6 pre (Namoroka)	2498	2443	1110.0	1127.6
Firefox 3.7 a1 pre (Minefield)	2573	2517	1013.8	1034.2

Screen for clarity:


Link to the results of the Peacekeeper test is one, so I didn’t litter the table - look here (Unknown unknown is Firefox 3.6 pre (Namoroka), for some reason the test didn’t recognize it).

findings

• As can be seen from the table of performance loss, or even in one case, the gain when using FBO is not more than 2.2%! We can say that FBO does not affect performance. But saves RAM. During the test, all versions of Firefox with FBO spent from 1 to 15 MB, when without it from 60 to 200 MB (!).
• New Opera 10.50 made a huge leap in comparison with its predecessor Oper'oy 10.10 and overtook even Firefox versions 3.6 and 3.7.
• But the negative dynamics of Oper'y 10.50 build 3199 (already the third build of version 10.50) is also noticeable compared to Oper'y 10.50 build 3172 (the very first build of version 10.50) - the results in the Peacekeeper test worsened: from 3400+ to 2800+ points (Test reports of the first build 10.50 I wrote here and here ).

Add. links

For those who do not know yet: “All versions and Russifications of Firefox” and similar storage from Oper'y (only there are no pre-alpha builds).

ZY Do not kick much, this is my first topic on Habré :)

Thanks for attention!

UPD: If someone added his Chrome 4.0.249.64 to the Peacekeeper test, let me add the fourth build of Opera 10.50 build 3202 released today. The result is 3688. Opera not only returned its original result, but also improved it, catching up with Chrome.