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

Re: [Public WebGL] Best practices questions




On Jan 10, 2011, at 10:26 AM, Thatcher Ulrich wrote:

> On Fri, Jan 7, 2011 at 5:45 PM, Chris Marrin <cmarrin@apple.com> wrote:
>> 
>> On Jan 7, 2011, at 7:21 AM, Thatcher Ulrich wrote:
>> 
>>> Re capabilities, have a look at webgl-bench.appspot.com ; I'm collecting this kind of info and showing aggregates of it.  The breakdowns are coarse but I think can be made finer as more data is accumulated.
>> 
>> This is a nice page, thanks for doing it. It would be great if you published some qualitative tables, saying "this combination of graphics, hardware and browser succeeds in these areas and fails in these". It would give the browser vendors good data on a wider variety of configurations than we have on hand.
> 
> By "qualitative", do you mean I should write up my interpretations of
> where I think the trouble spots are?  I.e. from looking at the data so
> far, I conclude that Firefox is the fastest webgl browser, followed by
> Webkit, followed by Chrome.  ANGLE doesn't expose Vertex Texture
> Fetch, which is unfortunate for anyone hoping for that feature.  Intel
> GPUs seem pretty weak relative to ATI and NVIDIA.  Some of the GPU
> data is a bit harder to interpret because ANGLE subsumes other GPUs
> (maybe only Intel at the moment?).  Also as we discussed earlier,
> Firefox doesn't expose a GPU string so the speed comparisons across
> browser are perhaps unfair, although it does come out ahead of the
> other browsers on my personal machine.

I really just mean sorting the data by graphics/hardware/browser and say "this combo passes these tests and fails these". It would help us understand if the problem is with some common library (like ANGLE), a particular graphics card, a given browser, etc.

-----
~Chris
cmarrin@apple.com




-----------------------------------------------------------
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
-----------------------------------------------------------