It's an asymmetric world

Page 1 of 2

Background

I've read reviews on the speed of Javascript on different browsers, and most of them have been done on MS Windows, a few on Mac OS X and several flavors of Linux. So I got curious and tried benchmarking the browsers I got installed on my old (but trusty) iBook G4 which is running Mac OS X 10.4 (Tiger).

For this informal series of tests I used three of the most popular (and often quoted) benchmarks:

SunSpider, Javascript Benchmark.
V8 Benchmark Suite (version 1).
Mozilla's Dromaeo.

All the tests were executed with the web browser being the only running application, to avoid background bias in the performance.

My laptop configuration

OS	Mac OS X, Version 10.4.11
Machine Model	PowerBook6,3
Memory	640 MB DDR SDRAM
CPU Type	PowerPC G4 (3.3)
Number Of CPUs	1
CPU Speed	800 MHz
L2 Cache	256KB
Bus Speed	133 MHz

Results

Web Browser	SunSpider Javascript Benchmark	SunSpider ratio [a]	Mozilla Dromaeo v2	Dromaeo ratio [a]	V8 Benchmark Suite - Version 1	V8 ratio [b]
Firefox 3.0.3	17862.6	1.00	56785.8	1.00	36	1.00
Firefox 3.1b1	15315.8	0.86	40458.8	0.71	40	0.90
Flock 2.0b3	35760.6	2.00	57197.2	1.01	33	1.09
Opera 9.52	49726.2	2.78	163440.4	2.88	27	1.33
Safari 3.1.2	48605.2	2.72	49492.6	0.87	30	1.20
Shiira 2.2	26385.4	1.48	82699.0	1.46	15	2.40
Sunrise 1.7.5	23684.8	1.33	83579.6	1.47	17	2.12
Camino 1.6.4	120023.4	12.15	NA [c]	NA [c]	10	3.60

[a] Ratio: time(web browser) / time(Firefox 3.0.3)
[b] Ratio: score(Firefox 3.0.3) / score(web browser)
[c] Could not finish the test, web browser generated continuous script timeouts.

Due to some problems with Camino, as well as the over one order of magnitude difference with Mozilla Firefox 3.0.3 in the SunSpider test, I have not included that web browser in the following plots.

Results from SunSpider tests
SunSpider test results