>+movd        m0,     [r2]
>+movd        m1,     [r3]
>+pmovzxbw    m2,     m0
>+pmovzxbw    m3,     m1
>+
>+movd        m0,     [r2 + r4]
>+movd        m1,     [r3 + r5]
>+pmovzxbw    m4,     m0
>+pmovzxbw    m5,     m1
>+
>+movd        m0,     [r2 + 2 * r4]
>+movd        m1,     [r3 + 2 * r5]
>+pmovzxbw    m6,     m0
>+pmovzxbw    m7,     m1
>+
>+psubw       m2,     m3
>+psubw       m4,     m5
>+psubw       m6,     m7
>+
>+lea         r2,     [r2 + 2 * r4]
>+lea         r3,     [r3 + 2 * r5]
>+
>+movd        m0,     [r2 + r4]
>+movd        m1,     [r3 + r5]
>+pmovzxbw    m3,     m0
>+pmovzxbw    m5,     m1
>+
>+psubw       m3,     m5
>+
>+movh    [r0],             m2
>+movh    [r0 + r1],        m4
>+movh    [r0 + 2 * r1],    m6
>+
>+lea     r0,               [r0 + 2 * r1]
>+
>+movh    [r0 + r1],        m3
>+
>+RET

un-debugged example
movd        m0,     [r2]
movd        m1,     [r3]
movd        m2,     [r2 + r4]
movd        m3,     [r3 + r5]
punpckldq m0, m2
punpckldq m1, m3
pmovzxbw m0, m0
pmovzxbw m1, m1
psubw m0, m1
movlps [r0], m0
movhps [r0 + r1], m0
 
_______________________________________________
x265-devel mailing list
[email protected]
https://mailman.videolan.org/listinfo/x265-devel

Reply via email to