[Mesa-dev] r600g: status of my work on the shader optimization

[Vadim Girlin]

On 02/22/2013 04:35 PM, Henri Verbeet wrote:
> For what it's worth, I get a number of failures in the Wine d3d9 tests
> on Cayman with this branch (9219c54b5c5a65c269124637a6e654eda4cdbb8e).
> I haven't looked at why yet.
>

Unfortunately I have no cayman card to reproduce the issues. I can only 
test on evergreen, and I don't see any problems with d3d9 tests. On 
evergreen all known problems are fixed already, and I have to rely on 
the help of the users to fix remaining issues with other chips.

If you'd like to help me with debugging the issues on cayman, then 
please read the "regression debugging" section in the r600-sb branch 
notes [1] (or possibly more up-to-date source here - [2]), it explains 
how to find the exact shader that causes regression. After locating the 
guilty shader, you only need to prepend

     R600_DUMP_SHADERS=2 R600_SB_DUMP=3

to the command line to produce the full dump for that shader, then 
please send it to me, and I'll do my best to fix the issue.

Thanks for testing this branch.

Vadim

[1]: http://people.freedesktop.org/~vadimg/r600-sb.html

[2]: 
http://cgit.freedesktop.org/~vadimg/mesa/plain/src/gallium/drivers/r600/sb/notes.markdown?h=r600-sb