[Mplayer-users] [mplayer PATCH] Probably final fix of fastmemcpy

[Arpi]

Hi,

> I want to suggest new version of fastmemcpy.
> What is noticeable:
> 1. MMX only support was added. It really speedups memory copying no Pentium-MMX, Celeron-1, P2 at least 
Great!

> (see don't applied previous patch)
This code is maintained by Felix, maybe he found it non-workable or slower.
I'm busy now with putting together the stable release version.

Btw there are some more slow C code in mplayer, which should be optimized
in MMX, the OSD/SUB renderer (libvo/osd.c) and YUV scaler (libvo/fsdga.c).
Can you look at them?
For OSD renderer, I want to change character bitmaps to be 16 byte aligned
and width%8==0, so MMX code will be much easier.


A'rpi / Astral & ESP-team

--
mailto:arpi at thot.banki.hu
http://esp-team.scene.hu

_______________________________________________
Mplayer-users mailing list
Mplayer-users at lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/mplayer-users