DOM access speed mini BenchMark : Safari vs FireFox vs Opera

[Jorge]

Based on the table shuffling code of the recent cljs thread "Randomize
HTML Table Rows from JavaScript " (http://tinyurl.com/56g37t), I get
this results :

e.replaceChild() (SAM's code) :

FF2 : 1x
FF3: 1.3x
Opera 9.5 : 4.8x
Safari r34469 : 8.4x


e.insertBefore() (Thomas' code) :

FF2 : 1x
FF3: 1.2x
Opera 9.5 : 4.7x
Safari r34469 : 10.8x

e.appendChild() (Jorge's code) :

FF2 : 1x
FF3: 1.1x
Opera 9.5 : 4.3x
Safari r34469 : 10.2x

As the JavaScript core (Sun Spider : webkit.org/perf/sunspider-0.9/
sunspider.html) benchmarks were already showing, Safari is the
fastest, but I wasn't expecting such **huge** differences in DOM
manipulation speeds...

The data : (got it on my Mac/OS 10.5.3)

FF2.0.0.14

159.1/100 ms (SAM)
104.2/100 ms (Thomas)
88.66/100 ms (Jorge)

FF3.0

123.23/100 ms (SAM)
86.7/100 ms (Thomas)
83.72/100 ms (Jorge)

Opera 9.5

33.25/100 ms (SAM)
22.33/100 ms (Thomas)
20.61/100 ms (Jorge)

Safari/WebKit r34469 :

18.92/100 ms (SAM)
9.69/100 ms (Thomas)
8.72/100 ms (Jorge)

Take it with a grain of salt.

--Jorge.

 

[Gregor Kofler]

Jorge meinte:

Based on the table shuffling code of the recent cljs thread "Randomize
HTML Table Rows from JavaScript " (http://tinyurl.com/56g37t), I get
this results :

As the JavaScript core (Sun Spider : webkit.org/perf/sunspider-0.9/
sunspider.html) benchmarks were already showing, Safari is the
fastest, but I wasn't expecting such **huge** differences in DOM
manipulation speeds...

The data : (got it on my Mac/OS 10.5.3)

FF2.0.0.14

159.1/100 ms (SAM)
104.2/100 ms (Thomas)
88.66/100 ms (Jorge)

FF3.0

123.23/100 ms (SAM)
86.7/100 ms (Thomas)
83.72/100 ms (Jorge)

Opera 9.5

33.25/100 ms (SAM)
22.33/100 ms (Thomas)
20.61/100 ms (Jorge)

Safari/WebKit r34469 :

18.92/100 ms (SAM)
9.69/100 ms (Thomas)
8.72/100 ms (Jorge)

Take it with a grain of salt.

Indeed. (WinXP, "normally" running system)

FF3 (all sorts of extensions activated): 94/64/63
Safari 3: 21/15/15
Opera 9.5: 15/0.5/0 (!)

Looking at the result I can state, that Opera is approx. 1,000,000,000
faster than Safari... Man this *is* fast.

Gregor

 

[Jorge]

Opera 9.5: 15/0.5/0 (!)

Looking at the result I can state, that Opera is approx. 1,000,000,000
faster than Safari... Man this *is* fast.

Yeah ! That last 0 means that that Opera of yours can do infinite work
in no time at all. Keep it in a safe place, that's worth millions.
Unfortunately mine read 90/37/32 and so it's not worth so much.

--Jorge.

 

[Gregor Kofler]

Jorge meinte:

Yeah ! That last 0 means that that Opera of yours can do infinite work
in no time at all. Keep it in a safe place, that's worth millions.

It covers CD-ROMs with a not-so-thin golden layer, too. It's so amazing.
The 0.5 mark isn't that bad either. I'll use it for brute-force cracking
of 256-bit keys.

Unfortunately mine read 90/37/32 and so it's not worth so much.

Perhaps some nightly build, where they left out the good parts.

Gregor