[cairo] cairo/glitz sample code

[Peter Clifton]

Back on the cairo list...
On Wed, 2008-07-23 at 10:20 +0530, Atharva wrote:
> Hi Peter,
> 		Thanks for your patch. I applied that to my cairo library and
> performance of sample programs improved considerably. Here are some
> interesting observations that I figured out. 
> 
> I used cairo-gears example as reference with  -glx option. 
> There are few sample tests associated with each back-end.
> 
> 
> Here is the analysis of various test cases:

Reformatted to ASCII-art:
 _____________________________________________________
|       | s/w mesa x86 | s/w mesa arm | opengl h/w    |
| TEST  |--------------+--------------+---------------|
|       |BEFORE: AFTER |BEFORE: AFTER |BEFORE: AFTER  |
|-------+------+-------+------+-------+------+--------|
|TRAP   | *N   : *Y    | *N   : *Y    |  Y   :  Y     |
|GRAD   | *Y   : *Y, I |  Y   :  Y    |  Y   :  Y     |
|COMP   |  N   :  N    |  N   :  N    |  Y   :  Y     |
|TEXT   |  N   :  N    |  N   :  N    |  Y   :  Y     |
|TEXT2  | &Y   : &N    |  N   :  N    |  Y   :  Y     |
|SHADOW |  N   :  N    |  N   :  N    |  Y   :  Y     |
|_______|______:_______|______:_______|______:________|


I was testing against the git version of the glitz library.

Just re-testing, I now get a failure in "TEXT2", which I don't recall
having had before. Also, "SHADOW" isn't working properly, whereas I
recall it did before.

I'm not sure quite what has changed, but I think mesa / Intel Xorg
drivers have been "upgraded" since I originally wrote the patch.

> 
> >From above observations we can draw few conclusions:
> 1. Glitz backend works well with opengl acccelerated hardware.
> 2. After applying patch some tests worked as  well as performace also
>    improved.
> 3. However, some test cases, which were working before, failed after
>    applying the patch.
> 
> Can anybody throw more light on this issue ? I need to verify whether
> my observations are correct. Thanks in advance.

I'm surprised you've got everything appearing to work with GL, before
the patch. I never tested explicitly with s/w mesa (I guess it was only
used for fallbacks), my Intel 945GM wouldn't work for most of the above
tests until the patch was applied.

What graphics card / driver are you testing with?

-- 
Peter Clifton

Electrical Engineering Division,
Engineering Department,
University of Cambridge,
9, JJ Thomson Avenue,
Cambridge
CB3 0FA

Tel: +44 (0)7729 980173 - (No signal in the lab!)