Re: clutter batching

[Emmanuele Bassi <ebassi gmail com>]

Hi;

Test-state is a bit of a edge case: it's meant to be the model used to stress test Clutter (and Cogl). In reality, most of what it does is wrong and deprecated, but it provides us with a good baseline for the worst case scenario of Clitter usage. I don't regard the interactive test suite as a performance indicator, or as code that ought to be used by third party developers.

Ciao,

 Emmanuele.

On Tuesday, 19 February 2013, mahesh gawali wrote:

Salmokar,

I followed up with your post on COGL mailing list since I was also curious about cogl-batching.

Here's  a response from Robert that might help you.

http://lists.freedesktop.org/archives/cogl/2013-February/000956.html

--

Mahesh

---

Date: Wed, 13 Feb 2013 15:35:29 +0530
Subject: clutter batching
From: salmokarjames gmail com
To: clutter-list gnome org

Hello,

While running the ./test-state test within clutter, and exporting COGL_DEBUG=batching , i am getting a lot of "len=1" output..

but within this bulk of "len=1" i also get this once in a while..

journal len = 24

clip stack batch len = 24

 vbo offset batch len =24

and some times len = 48 , 72 etc..

so does that mean that batching is not being performed efficiently(or at all) because majority of the times the len =1?

and why the len varies from 1 to 24 to 72 once in a while? what does this actually mean? and is this test case suitable for batching..?

I am new to clutter development and help and explanation would benefit.

Thanks,

Kind regards,

Salmokar.

_______________________________________________ clutter-list mailing list clutter-list gnome org https://mail.gnome.org/mailman/listinfo/clutter-list

-- 
W: http://www.emmanuelebassi.name
B: http://blogs.gnome.org/ebassi/