WOW IE9, you’ve come a long way…
Posted on September 16, 2010 by Chris Koenig
Nice job IE9 team! The animation was pretty smooth, but stuttered a bit at 26 and 69. This is still WAY better than the scores were for IE8…
I also ran the SunScript JavaScript benchmark and compared it to a similar run on Google Chrome 6.0.472.59:
TEST COMPARISON FROM (IE9) TO (Chrome) DETAILS
=============================================================================
** TOTAL **: 1.15x as fast 934.4ms +/- 5.6% 815.4ms +/- 7.7% significant
=============================================================================
3d: - 157.6ms +/- 26.0% 119.4ms +/- 9.0%
cube: ?? 40.4ms +/- 10.3% 41.8ms +/- 11.0% not conclusive: might be *1.03x as slow*
morph: 1.43x as fast 58.6ms +/- 12.5% 41.0ms +/- 18.6% significant
raytrace: - 58.6ms +/- 65.8% 36.6ms +/- 7.8%
access: - 101.8ms +/- 15.6% 86.0ms +/- 12.6%
binary-trees: 3.27x as fast 19.6ms +/- 65.4% 6.0ms +/- 65.5% significant
fannkuch: *1.36x as slow* 29.4ms +/- 7.1% 40.0ms +/- 31.3% significant
nbody: 1.57x as fast 44.8ms +/- 5.3% 28.6ms +/- 5.0% significant
nsieve: *1.43x as slow* 8.0ms +/- 11.0% 11.4ms +/- 31.4% significant
bitops: *1.43x as slow* 51.8ms +/- 3.6% 74.0ms +/- 21.9% significant
3bit-bits-in-byte: ?? 4.6ms +/- 24.2% 5.8ms +/- 35.2% not conclusive: might be *1.26x as slow*
bits-in-byte: *1.58x as slow* 12.8ms +/- 4.3% 20.2ms +/- 34.1% significant
bitwise-and: *2.09x as slow* 9.4ms +/- 7.2% 19.6ms +/- 8.5% significant
nsieve-bits: *1.14x as slow* 25.0ms +/- 3.5% 28.4ms +/- 33.3% significant
controlflow: *1.58x as slow* 6.6ms +/- 16.8% 10.4ms +/- 131.4% significant
recursive: *1.58x as slow* 6.6ms +/- 16.8% 10.4ms +/- 131.4% significant
crypto: ?? 53.0ms +/- 8.6% 57.4ms +/- 20.7% not conclusive: might be *1.08x as slow*
aes: *1.23x as slow* 20.2ms +/- 8.0% 24.8ms +/- 52.1% significant
md5: - 16.6ms +/- 15.5% 15.6ms +/- 16.5%
sha1: ?? 16.2ms +/- 3.4% 17.0ms +/- 30.6% not conclusive: might be *1.05x as slow*
date: 1.30x as fast 116.6ms +/- 8.3% 89.8ms +/- 13.8% significant
format-tofte: 1.34x as fast 58.2ms +/- 12.9% 43.4ms +/- 32.6% significant
format-xparb: 1.26x as fast 58.4ms +/- 13.8% 46.4ms +/- 11.8% significant
math: *1.23x as slow* 78.2ms +/- 5.3% 96.4ms +/- 19.9% significant
cordic: *12.7x as slow* 3.0ms +/- 0.0% 38.2ms +/- 50.5% significant
partial-sums: 1.16x as fast 50.2ms +/- 8.1% 43.2ms +/- 19.9% significant
spectral-norm: 1.67x as fast 25.0ms +/- 6.1% 15.0ms +/- 41.4% significant
regexp: 2.18x as fast 68.6ms +/- 32.0% 31.4ms +/- 11.4% significant
dna: 2.18x as fast 68.6ms +/- 32.0% 31.4ms +/- 11.4% significant
string: 1.20x as fast 300.2ms +/- 4.2% 250.6ms +/- 9.0% significant
base64: ?? 23.8ms +/- 42.3% 27.6ms +/- 17.0% not conclusive: might be *1.16x as slow*
fasta: 1.41x as fast 53.4ms +/- 5.1% 37.8ms +/- 28.4% significant
tagcloud: - 83.0ms +/- 20.1% 76.8ms +/- 40.5%
unpack-code: 1.24x as fast 89.2ms +/- 4.9% 71.8ms +/- 5.4% significant
validate-input: 1.39x as fast 50.8ms +/- 10.0% 36.6ms +/- 9.2% significant
Overall, the JavaScript engine in the IE9 Beta was still slower than Chrome, but not by as much as I expected. Chrome was only 1.15x faster than IE9 Beta, and in some cases was slower than IE by as much as 2x-3x! It looks like the test will tag any result over 1.1x as “significant” from a comparison perspective, but 1.15x doesn’t feel “significant” – more like “slightly”, but I guess I’m a little biased…
Now that IE supports HTML 5, I guess it’s time for me to finally get around to learning more about it. I’m sort of pre-occupied with Silverlight and Windows Phone 7 (have you noticed) but I had to do this special shout-out to the IE team for a job well done!
Comments (2)








You are overstating the numbers in the javascript benchmark, as are they.
934.4/815.4 = 1.15. That doesn’t mean it’s 1.15x faster. It means IE9 spent 1.15x the time Chrome spent. It’s only 0.15x slower, not 1.15x slower.
In the math section, Chrome loses to IE due to the one sub-component: cordic. I’d like to know more information about that. Chrome soundly trounces it in the other 2 sub-components. 12.7x is quite a difference.
I got this piece of shit IE-9 through windows update. Why would they send out something you can not usse with everyday use. My bank tells me they can not fix the problem as it is not certified. So is short, I can not use my on-line banking. This is a shit until it gets certified. Way to Go Microsoft!