sub8x8_dct_dc                   (16-bit)
sub8x16_dct_dc                  (16-bit)
optimize_chroma_2x4_dc          (8 and 16-bit) [Warning: may be hard]
predict_8x16_p                  (8 and 16 bit) (claimed)
sub8x8_dct8                     (16-bit) [Warning: may be hard]
add8x8_idct8                    (16-bit) [Warning: may be hard]
sub16x16_dct8                   (16-bit) [Warning: may be hard]
add16x16_idct8                  (16-bit) [Warning: may be hard]
>mmx versions of SA(T|8)D       (16-bit) [Warning: may be hard]
>mmx version of pixel_avg2_w8   (16-bit)
>mmx version of sub8x8_dct      (16-bit)
intra_sad 8x8c                  (16-bit) (claimed)
intra_satd 8x8                  (16-bit) [Warning: may be hard]
intra_sad 16x16                 (16-bit) (claimed, but haven't seen the person around)
intra_satd 8x16c                (8 and 16-bit)
intra_sad 8x16c                 (8 and 16-bit)
optimize_chroma_2x2_dc          (16-bit) [Warning: may be hard]
Make any function significantly faster (any)

x262 functions:

sub8x8_dct_mpeg2
add8x8_idct_mpeg2
x264_coeff_level_run64
x264_coeff_level_run63
hpel_filter_mpeg2
dequant_mpeg2_inter
mc_chroma_mpeg2
get_ref_mpeg2
dequant_mpeg2_intra
mc_luma_mpeg2
zigzag_scan_8x8_field_mpeg2