Yes, it seems you're right ! ;)
Guillaume -------- Message d'origine-------- De: Nikita Egorov [mailto:[EMAIL PROTECTED] Date: jeu. 03/07/2008 10:36 À: GARDET Guillaume Cc: Matteo Fortini; directfb-dev@directfb.org Objet : Re: RE : [directfb-dev] carmine + PPC with carmine_fb > "changing of wrong order of bytes on-the-fly" is the easy way but it will > use more the CPU. Do you mean CPU loading? It is not quite right. For example, in my case with PowerPC architecture I have replaced only _one_ CPU command ( Load Word => Load Word Byte Reverse) . Thus, there is no difference for CPU at all ! Second,the memcpy() function is working without any changing, as is ! Third, in assembler code of real function processing gfx data the "Store/Load" commands are seldom and adding one command ( as in last case ) make worse the speed of work only a few percent or less.. -- Best Regards Nikita Egorov [EMAIL PROTECTED] [EMAIL PROTECTED]
_______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
