| Branch: | Revision:

ffmpeg / libavcodec / i386 / dsputil_mmx_rnd.h @ f259747b

History | View | Annotate | Download (15.7 KB)

# Date Author Comment
0c1a9eda 02/11/2003 04:35 PM Zdenek Kabelac
  • UINTX -> uintx_t INTX -> intx_t

Originally committed as revision 1578 to svn://

88e9ab1f 01/09/2003 10:10 PM Michael Niedermayer

typo (noticed by kabi)

Originally committed as revision 1427 to svn://

db794953 01/09/2003 08:42 PM Michael Niedermayer

qpel fix

Originally committed as revision 1426 to svn://

5a508a98 01/07/2003 01:00 PM Michael Niedermayer

PIC / ebx fix

Originally committed as revision 1408 to svn://

0b093b6f 01/07/2003 11:59 AM Michael Niedermayer

more gcc bug workarounds

Originally committed as revision 1407 to svn://

3df6fa77 01/07/2003 11:07 AM Michael Niedermayer

gcc 3.2.2 -O3 bug workaround (older gcc are very likely affected too but didnt check)

Originally committed as revision 1406 to svn://

3178ee4c 01/05/2003 08:59 PM Michael Niedermayer

qpel in b frames bugfixes

Originally committed as revision 1398 to svn://

826f429a 01/05/2003 03:57 PM Michael Niedermayer

qpel in mmx2/3dnow
qpel refinement quality parameter

Originally committed as revision 1393 to svn://

b3184779 09/11/2002 12:39 PM Michael Niedermayer

fixing 2 small qpel bugs

Originally committed as revision 915 to svn://

9f361534 08/20/2002 07:23 PM Michael Niedermayer

dont trash mm0, its still used in the next iteration of the loop (fixes strange horizontal lines on some blocks in b frames on mmx only cpus)

Originally committed as revision 855 to svn://

d3cd0742 06/03/2002 05:31 PM Zdenek Kabelac
  • implemented remaing avg_ pixel functions (these are not used offen)

Originally committed as revision 656 to svn://

fca0f0e5 05/30/2002 03:14 PM Zdenek Kabelac
  • removed MANGLE from macros for setting constants
  • using MOVQ_WONE/MOVQ_BFE as two instruction instead of static memory value access
    as its always faster
  • PAVGB_MMX macro is using now mm6 -> mm7 is unmodified
  • replaced original pixels_xy2_mmx with new faster and equal implementation...
ff965706 05/29/2002 08:00 PM Zdenek Kabelac
  • using unrolled loops seems to help to MMX Celerons
  • calculating average in parallel to use two MMX units
  • disabled xy2 put code - it's producing rouding errors

Originally committed as revision 626 to svn://

91abb473 05/29/2002 05:16 PM Zdenek Kabelac
  • code with new PAVGB for MMX only CPU splited into separate file
    and being compiled in the same way as _avg.h
  • PAVG_MMX macros accept also output parameter
  • implemented faster put_pixels_xy2, but it has slightly smaller precission.
    But there is not visible difference in the image quality - might be eventualy...