[x264-devel] NOTE: Altivec/PPC port is broken

Guillaume POIRIER poirierg at gmail.com
Thu Mar 20 10:19:14 CET 2008


Hello folks,

The different changes that have been made to the original C code lead
Altivec routines to stop matching the output of C code.
Here's what checkasm has to say about it:

./tools/checkasm
x264: using random seed 3183971084
x264: ALTIVEC against C
 - pixel sad :           [OK]
  - pixel ssd :           [OK]
  - pixel satd :          [OK]
 - pixel sa8d :          [OK]
 - pixel sad_x3 :        [OK]
 - pixel sad_x4 :        [OK]
 - sub_dct4 :            [OK]
 - sub_dct8 :            [OK]
 - add_idct4 :           [OK]
 - add_idct8 :           [OK]
scan_4x4 [FAILED]
scan_4x4ac [FAILED]
 - zigzag_frame :        [FAILED]
scan_4x4 [FAILED]
scan_4x4ac [FAILED]
 - zigzag_field :        [FAILED]
  - mc luma :             [OK]
  - mc chroma :           [OK]
hpel filter differs at plane c line 0
9cffd726 bfd55fbd 45009a90 7b957b85 adab4a77 c7ff7e16 f0a251bb
d86654d3 562abecf a9787582 496b7aa3 855f2714
9bffd726 bfd55fbd 45009a90 7b957b85 adab4a77 c7ff7e16 f0a251bb
d86654d3 562abecf a9787582 496b7aa3 855c3300
  - hpel filter :         [FAILED]
  - intra pred :          [OK]
  - deblock :             [OK]
  - quant :               [OK]
  - dequant :             [OK]
x264: at least one test has failed. Go and fix that Right Now!

I shall fix it as soon as I can. Meanwhile, please stick to pre-GIT
versions of x264, which still work IIRC.

Guillaume
-- 
A soldier will fight long and hard for a bit of colored ribbon.
 -- Napoleon Bonaparte


More information about the x264-devel mailing list