swscale/swscale_unscaled : add X86_64 (SSE2 and AVX) for uyvyto422

and checkasm test

2 files changed, 169 insertions, 0 deletions

diff --git a/libswscale/x86/rgb2rgb.c b/libswscale/x86/rgb2rgb.c
index e5f318a..1191081 100644
--- a/libswscale/x86/rgb2rgb.c
+++ b/libswscale/x86/rgb2rgb.c
@@ -150,6 +150,15 @@ void ff_shuffle_bytes_1230_ssse3(const uint8_t *src, uint8_t *dst, int src_size)
 void ff_shuffle_bytes_3012_ssse3(const uint8_t *src, uint8_t *dst, int src_size);
 void ff_shuffle_bytes_3210_ssse3(const uint8_t *src, uint8_t *dst, int src_size);
 
+#if ARCH_X86_64
+void ff_uyvytoyuv422_sse2(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
+                          const uint8_t *src, int width, int height,
+                          int lumStride, int chromStride, int srcStride);
+void ff_uyvytoyuv422_avx(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
+                         const uint8_t *src, int width, int height,
+                         int lumStride, int chromStride, int srcStride);
+#endif
+
 av_cold void rgb2rgb_init_x86(void)
 {
     int cpu_flags = av_get_cpu_flags();
@@ -167,6 +176,11 @@ av_cold void rgb2rgb_init_x86(void)
         rgb2rgb_init_avx();
 #endif /* HAVE_INLINE_ASM */
 
+    if (EXTERNAL_SSE2(cpu_flags)) {
+#if ARCH_X86_64
+        uyvytoyuv422 = ff_uyvytoyuv422_sse2;
+#endif
+    }
     if (EXTERNAL_SSSE3(cpu_flags)) {
         shuffle_bytes_0321 = ff_shuffle_bytes_0321_ssse3;
         shuffle_bytes_2103 = ff_shuffle_bytes_2103_ssse3;
@@ -174,4 +188,9 @@ av_cold void rgb2rgb_init_x86(void)
         shuffle_bytes_3012 = ff_shuffle_bytes_3012_ssse3;
         shuffle_bytes_3210 = ff_shuffle_bytes_3210_ssse3;
     }
+    if (EXTERNAL_AVX(cpu_flags)) {
+#if ARCH_X86_64
+        uyvytoyuv422 = ff_uyvytoyuv422_avx;
+#endif
+    }
 }
diff --git a/libswscale/x86/rgb_2_rgb.asm b/libswscale/x86/rgb_2_rgb.asm
index db45e31..156b4d2 100644
--- a/libswscale/x86/rgb_2_rgb.asm
+++ b/libswscale/x86/rgb_2_rgb.asm
@@ -32,6 +32,16 @@ pb_shuffle3210: db 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12
 
 SECTION .text
 
+%macro RSHIFT_COPY 3
+; %1 dst ; %2 src ; %3 shift
+%if cpuflag(avx)
+    psrldq  %1, %2, %3
+%else
+    mova           %1, %2
+    RSHIFT         %1, %3
+%endif
+%endmacro
+
 ;------------------------------------------------------------------------------
 ; shuffle_bytes_## (const uint8_t *src, uint8_t *dst, int src_size)
 ;------------------------------------------------------------------------------
@@ -84,3 +94,143 @@ SHUFFLE_BYTES 0, 3, 2, 1
 SHUFFLE_BYTES 1, 2, 3, 0
 SHUFFLE_BYTES 3, 0, 1, 2
 SHUFFLE_BYTES 3, 2, 1, 0
+
+;-----------------------------------------------------------------------------------------------
+; uyvytoyuv422(uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
+;              const uint8_t *src, int width, int height,
+;              int lumStride, int chromStride, int srcStride)
+;-----------------------------------------------------------------------------------------------
+%macro UYVY_TO_YUV422 0
+cglobal uyvytoyuv422, 9, 14, 8, ydst, udst, vdst, src, w, h, lum_stride, chrom_stride, src_stride, wtwo, whalf, tmp, x, back_w
+    pxor         m0, m0
+    pcmpeqw      m1, m1
+    psrlw        m1, 8
+
+    movsxdifnidn            wq, wd
+    movsxdifnidn   lum_strideq, lum_strided
+    movsxdifnidn chrom_strideq, chrom_strided
+    movsxdifnidn   src_strideq, src_strided
+
+    mov     back_wq, wq
+    mov      whalfq, wq
+    shr      whalfq, 1     ; whalf = width / 2
+
+    lea srcq, [srcq + wq * 2]
+    add    ydstq, wq
+    add    udstq, whalfq
+    add    vdstq, whalfq
+
+.loop_line:
+    mov          xq, wq
+    mov       wtwoq, wq
+    add       wtwoq, wtwoq ; wtwo = width * 2
+
+    neg       wq
+    neg    wtwoq
+    neg   whalfq
+
+    ;calc scalar loop count
+    and       xq, mmsize * 2 - 1
+    je .loop_simd
+
+    .loop_scalar:
+        mov             tmpb, [srcq + wtwoq + 0]
+        mov [udstq + whalfq], tmpb
+
+        mov             tmpb, [srcq + wtwoq + 1]
+        mov     [ydstq + wq], tmpb
+
+        mov             tmpb, [srcq + wtwoq + 2]
+        mov [vdstq + whalfq], tmpb
+
+        mov             tmpb, [srcq + wtwoq + 3]
+        mov [ydstq + wq + 1], tmpb
+
+        add      wq, 2
+        add   wtwoq, 4
+        add  whalfq, 1
+        sub      xq, 2
+        jg .loop_scalar
+
+    ; check if simd loop is need
+    cmp      wq, 0
+    jge .end_line
+
+    .loop_simd:
+        movu    m2, [srcq + wtwoq             ]
+        movu    m3, [srcq + wtwoq + mmsize    ]
+        movu    m4, [srcq + wtwoq + mmsize * 2]
+        movu    m5, [srcq + wtwoq + mmsize * 3]
+
+        ; extract y part 1
+        RSHIFT_COPY    m6, m2, 1 ; UYVY UYVY -> YVYU YVY...
+        pand           m6, m1; YxYx YxYx...
+
+        RSHIFT_COPY    m7, m3, 1 ; UYVY UYVY -> YVYU YVY...
+        pand           m7, m1 ; YxYx YxYx...
+
+        packuswb       m6, m7 ; YYYY YYYY...
+        movu [ydstq + wq], m6
+
+        ; extract y part 2
+        RSHIFT_COPY    m6, m4, 1 ; UYVY UYVY -> YVYU YVY...
+        pand           m6, m1; YxYx YxYx...
+
+        RSHIFT_COPY    m7, m5, 1 ; UYVY UYVY -> YVYU YVY...
+        pand           m7, m1 ; YxYx YxYx...
+
+        packuswb                m6, m7 ; YYYY YYYY...
+        movu [ydstq + wq + mmsize], m6
+
+        ; extract uv
+        pand       m2, m1   ; UxVx...
+        pand       m3, m1   ; UxVx...
+        pand       m4, m1   ; UxVx...
+        pand       m5, m1   ; UxVx...
+
+        packuswb   m2, m3   ; UVUV...
+        packuswb   m4, m5   ; UVUV...
+
+        ; U
+        pand       m6, m2, m1 ; UxUx...
+        pand       m7, m4, m1 ; UxUx...
+
+        packuswb m6, m7 ; UUUU
+        movu   [udstq + whalfq], m6
+
+
+        ; V
+        psrlw      m2, 8  ; VxVx...
+        psrlw      m4, 8  ; VxVx...
+        packuswb   m2, m4 ; VVVV
+        movu   [vdstq + whalfq], m2
+
+        add   whalfq, mmsize
+        add    wtwoq, mmsize * 4
+        add       wq, mmsize * 2
+        jl .loop_simd
+
+    .end_line:
+        add        srcq, src_strideq
+        add        ydstq, lum_strideq
+        add        udstq, chrom_strideq
+        add        vdstq, chrom_strideq
+
+        ;restore initial state of line variable
+        mov           wq, back_wq
+        mov          xq, wq
+        mov      whalfq, wq
+        shr      whalfq, 1     ; whalf = width / 2
+        sub          hd, 1
+        jg .loop_line
+
+    RET
+%endmacro
+
+%if ARCH_X86_64
+INIT_XMM sse2
+UYVY_TO_YUV422
+
+INIT_XMM avx
+UYVY_TO_YUV422
+%endif