[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] WebKit 20-50% slower than Chrome/Minefield using large Float32Array Arrays



At 9:26 PM -0800 2/22/11, Vladimir Vukicevic wrote:
>----- Original Message -----
>> At 4:51 PM -0800 2/17/11, Kenneth Russell wrote:
>> >Just FYI, further optimizations are planned for typed arrays in V8.
>> >Currently certain optimizations are done for normal JavaScript arrays
>> >that aren't in place for typed arrays.
>>
>> I'll say!
>>
>> I just tested Chrome v10.0.648.82 and I'm getting 98 models steps per
>> second (without rendering) using regular arrays:
>> http://visual.local/seasons/earth/model2d-reg-arrays.html
>>
>> and only 37 model steps per second using JavaScript Typed arrays:
>> http://visual.local/seasons/earth/model2d.html
>>
>> However Chrome is blowing away Minefield: 4.0b12pre (2011-02-22)
>> Regular/Typed 13/12.2 and WebKit v79303 Regular/Typed: 24.7/14.5.
>
>I'm confused, what happened to "Minefield is reversed running about 41 fps using Typed and 28 fps using regular arrays"?  Did we regress something?

I've been updating the code and my computer is loaded differently at different times so without further investigation you should only draw conclusions from the relative differences reported at one time.

In my tests I was running only one WebGL-enabled browser at the same time and only one window open running the specific computational visualization.

I am also surprised by the differences -- I don't know an easy way to go backwards running Minefield or I'd try that first.
-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email:
unsubscribe public_webgl
-----------------------------------------------------------