Added support for CPU acceleration, RGBA float -> RGB u8 and RGBA float ->

Added support for CPU acceleration, RGBA float -> RGB u8 and RGBA float ->

2008-05-23  Jan Heller  <jheller@svn.gnome.org>

        Added support for CPU acceleration, RGBA float -> RGB u8 and
        RGBA float -> RGBA u8 SSE accelerated conversions.

        * babl/Makefile.am: Added babl-cpuaccel.[ch].
        * babl/babl-cpuaccel.h:
        * babl/babl-cpuaccel.c (babl_cpu_accel_get_support),
        (babl_cpu_accel_set_use), (arch_get_vendor), (arch_accel_intel),
        (arch_accel_amd), (arch_accel_centaur), (arch_accel_cyrix),
        (sigill_handler), (arch_accel_sse_os_support), (arch_accel),
        (cpu_accel): Runtime CPU detection code. Ported from GIMP.
        * babl/babl-internal.h: Include babl-cpuaccel.h.
        * babl/babl-memory.c (babl_malloc): Make babl_malloc align memory to
        BABL_ALIGN==16 boundaries.
        * babl/babl.c (babl_init): Enabled CPU acceleration.
        * configure.ac: Added compile time MMX/SSE/AltiVec detection. Ported
        from GIMP.
        * extensions/Makefile.am: Added SSE_EXTRA_CFLAGS for sse-fixups.c
        compilation.
        * extensions/sse-fixups.c (conv_rgbaF_linear_rgb8_linear),
        (conv_rgbaF_linear_rgba8_linear), (init): Added RGBA float -> RGB u8
        and RGBA float -> RGBA u8 SSE accelerated conversions.

svn path=/trunk/; revision=316