x86: Add high bitdepth cfl_ac_444 AVX2 asm

cfl_ac_444_w4_16bpc_c:      177.5
cfl_ac_444_w4_16bpc_ssse3:   16.1
cfl_ac_444_w4_16bpc_avx2:    16.0
cfl_ac_444_w8_16bpc_c:      268.3
cfl_ac_444_w8_16bpc_ssse3:   31.7
cfl_ac_444_w8_16bpc_avx2:    25.8
cfl_ac_444_w16_16bpc_c:     370.3
cfl_ac_444_w16_16bpc_ssse3:  50.8
cfl_ac_444_w16_16bpc_avx2:   40.5
cfl_ac_444_w32_16bpc_c:     841.9
cfl_ac_444_w32_16bpc_ssse3: 129.9
cfl_ac_444_w32_16bpc_avx2:   92.8