> > > The 2.6.10 davincifb is working fine. What kind of performance issues? :)
I have run df_dok ,, here are the results (without accel, Graphics/Software Rasterizer) But unfortunately these are not from the GIT ! I know the GIT is faster for RGB 16 bit.. i will shortly update with the numbers from GIT as well But atleast this gives a base for people to compare their Davinci performances Draw Operation DFB 1.1.1 / Davinci - no accel Anti-aliased Text 40.449 KChars/sec Anti-aliased Text (blend) 9.381 KChars/sec Fill Rectangle 85.836 MPixel/sec Fill Rectangle (blend) 2.302 MPixel/sec Fill Rectangles [10] 89.579 MPixel/sec Fill Rectangles [10] (blend) 2.317 MPixel/sec Fill Triangles 53.272 MPixel/sec Fill Triangles (blend) 2.246 MPixel/sec Draw Rectangle 4.333 KRects/sec Draw Rectangle (blend) 0.820 KRects/sec Draw Lines [10] 15.460 KLines/sec Draw Lines [10] (blend) 3.631 KLines/sec Fill Spans 70.012 MPixel/sec Fill Spans (blend) 2.274 MPixel/sec Blit 28.843 MPixel/sec Blit 180 17.675 MPixel/sec Blit colorkeyed 16.628 MPixel/sec Blit destination colorkeyed 6.448 MPixel/sec Blit with format conversion 13.702 MPixel/sec Blit with colorizing 3.558 MPixel/sec Blit from 32bit (blend) 2.915 MPixel/sec Blit from 32bit (blend) with colorizing 1.259 MPixel/sec Stretch Blit 22.205 MPixel/sec Stretch Blit colorkeyed 15.066 MPixel/sec But the question is, If we use accel how better can we get ? df_dok.c does not compile straight ... needs a small fix .. patch available here http://www.nabble.com/DirectFB-1.0.1-Examples-(update-needed-)-td14965126.html > The firmware sources will most likely be released under GPL soon. I'm not > sure > if we are allowed to send binaries built from Code Composer. > Dennis, Thats great news, So we will able to better these numbers ! -- Sriram Neelakandan Author - Embedded Linux System Design And Development ( http://tinyurl.com/2doosu)
_______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev