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

RE: [Public WebGL] JavaScript matrix libraries really 5x faster in Minefield than Chrome?



One more thing to consider is the quality of the code.

 

I used to use the J3DIMath.js used in http://www.khronos.org/webgl/wiki/Tutorial.  We found at least a bug after spending hours on it. Please see http://www.khronos.org/message_boards/viewtopic.php?f=46&t=3231&p=9862#p9862 discussion. I switched to mjs.

 

http://learningwebgl.com/cookbook/index.php/Main_Page#Matrix_libraries shows various matrix libraries.

 

I’d be great if there is a recommended high-quality and actively/well-maintained matrix library.

 

Best,

-- On Lee

 

From: owner-public_webgl@khronos.org [mailto:owner-public_webgl@khronos.org] On Behalf Of Gregg Tavares (wrk)
Sent: Monday, February 14, 2011 9:16 AM
To: public webgl
Subject: Re: [Public WebGL] _javascript_ matrix libraries really 5x faster in Minefield than Chrome?

 

So I probably shouldn't respond to this because being on the Chrome team no matter what I say it will sound defensive but:

 

*) Are you sure these are valid tests? Rumor is some _javascript_ engines will see that the results you are calculating are never used and optimize the entire test away. I don't know if Minefield goes that far though if it did that would be pretty cool

 

*) Are you sure they are computing the same things? For example mjs doesn't appear to have a true matrix inverse so you're comparing apples to oranges vs the other  libraries there.

 

*) I added the math library from the tdl stuff. They are slightly faster at some things and slower at others. Attached (sorry, new to git)

 

*) Matrix math isn't everything. For example running the webgl aquarium on FF4b11 vs Chromium 11.0.668.0 (74525) on your hardware (I have the same machine) using this url

 

At 50 fish

 

FF: 40fps

Chromium: 60fps

 

At 1000 fish

 

FF: 25fps

Chromium: 60fps

 

I bring that up because it's hard to compare matrix math libraries as they don't all do the same thing. For example, many of the matrix math libraries create objects like a Matrix4 object. Others don't and just use _javascript_ arrays or Float32Array completely un-augmented. Which is better? No idea. If it turns out objects are slower but more convenient which would you pick?  Another example, some libraries work better if you give them a destination variable. I think mjs this is true and I know the tdl fast library (which sadly is not so fast) also requires that. But your benchmark doesn't take those differences into account. Another is that you're almost always doing your math in place meaning you pass in a matrix and expect that matrix to be translated, rotated or scaled in place rather than returning a new matrix and leaving the original alone. That difference is not taken into account.


Of course none of that explains why FF is 5 times faster for your test as it is.

 

 

On Sun, Feb 13, 2011 at 8:55 PM, Stephen Bannasch <stephen.bannasch@deanbrook.org> wrote:

Are _javascript_ matrix libraries suitable for WebGL programming really 5x faster in Minefield than Chrome ... and about twice asfast in Chrome than in a WebKit nightly?

 

This is what I found with an adaptation of with Brandon Jones's benchmarks in his glmatrix library: https://glmatrix.googlecode.com/hg/

 

See the data below ...

 

I was working with the _javascript_ graphing library flotr and wanting to pick a matrix library for some webgl work so I adapted the benchmarks Brandon Jones created in his glmatrix library: https://glmatrix.googlecode.com/hg/.

 

My changes include:

- I'm only including the benchmarks.

- I've updated to the latest mjs as of Dec 15: 16:8e5b0944ef1e and included it in several more tests.

- I've also added a graph display of the results using flotr, see: http://solutoire.com/flotr/

 

My fork of these benchmarks are available here: https://github.com/stepheneb/webgl-matrix-benchmarks

 

And you can run them in a browser with webgl at this url: http://stepheneb.github.com/webgl-matrix-benchmarks/matrix_benchmark.html

 

[The somewhat lame reason this is a fork is that it was taking me longer thanI wanted using mercurial (don't use hg much) to follow the typical workflow I use on github, fork a repo, push my changes into a branch and send a pull request to the original developer ...]

 

The following data were generated on a MacBook Pro running MacOS X 10.6.6 with a 2.66 GHz Intel Core i7 and a NVIDIA GeForce GT330M 512MB

 

Google Chrome 9.0.597.102

 

 

 

WebKit nightly 78038

 

 

Minefield 4.0b12pre (2011-02-13)