[FFmpeg-devel] [PATCH] avcodec/mips: msa optimizations for vc1dsp

[gxw]

    >>+    TRANSPOSE4x4_SW_SW(in_l0, in_l1, in_l2, in_l3, t_l1, t_l2, t_l3, t_l4);
    >>+    TRANSPOSE4x4_SW_SW(in_r4, in_r5, in_r6, in_r7, in_l0, in_l1, in_l2, in_l3);
    >>+    TRANSPOSE4x4_SW_SW(in_l4, in_l5, in_l6, in_l7, in_l4, in_l5, in_l6, in_l7);
    >>+    in_r4 = t_l1, in_r5 = t_l2, in_r6 = t_l3, in_r7 = t_l4;
    >
    >It's better to transpose 'in_l0, in_l1, in_l2, in_l3' directly into themselves, and ' in_r4, in_r5, in_r6, in_r7' the same.
    >>+PUT_VC1_MSPEL_MC_MSA(2, 1);
    >>+PUT_VC1_MSPEL_MC_MSA(2, 2);
    >>+PUT_VC1_MSPEL_MC_MSA(2, 3);
    >>+
    >>+PUT_VC1_MSPEL_MC_MSA(3, 1);
    >>+PUT_VC1_MSPEL_MC_MSA(3, 2);
    >>+PUT_VC1_MSPEL_MC_MSA(3, 3);
    >
    >About the 'cnst_para*' used in put_vc1_mspel_mc_h_v_msa, maybe you can reference the usage of 'shift_value'.
    >It may reduce some if clause especially in 'ff_put_vc1_mspel_mc ## hmode ## vmode ## _16_msa'.
    
	Thx, will fix in v2.
    _______________________________________________
    ffmpeg-devel mailing list
    ffmpeg-devel at ffmpeg.org
    https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
    
    To unsubscribe, visit link above, or email
    ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".