[FFmpeg-devel] [PATCH v3 1/2] swscale/x86/output.asm: add x86-optimized planer gbr yuv2anyX functions
Mark Reid
mindmark at gmail.com
Mon Dec 13 01:12:25 EET 2021
On Wed, Nov 24, 2021 at 1:15 PM <mindmark at gmail.com> wrote:
> From: Mark Reid <mindmark at gmail.com>
>
> changes since v2:
> * fixed label
> changes since v1:
> * remove vex intruction on sse4 path
> * some load/pack marcos use less intructions
> * fixed some typos
>
> yuv2gbrp_full_X_4_512_c: 12757.6
> yuv2gbrp_full_X_4_512_sse2: 8946.6
> yuv2gbrp_full_X_4_512_sse4: 5138.6
> yuv2gbrp_full_X_4_512_avx2: 3889.6
> yuv2gbrap_full_X_4_512_c: 15368.6
> yuv2gbrap_full_X_4_512_sse2: 11916.1
> yuv2gbrap_full_X_4_512_sse4: 6294.6
> yuv2gbrap_full_X_4_512_avx2: 3477.1
> yuv2gbrp9be_full_X_4_512_c: 14381.6
> yuv2gbrp9be_full_X_4_512_sse2: 9139.1
> yuv2gbrp9be_full_X_4_512_sse4: 5150.1
> yuv2gbrp9be_full_X_4_512_avx2: 2834.6
> yuv2gbrp9le_full_X_4_512_c: 12990.1
> yuv2gbrp9le_full_X_4_512_sse2: 9118.1
> yuv2gbrp9le_full_X_4_512_sse4: 5132.1
> yuv2gbrp9le_full_X_4_512_avx2: 2833.1
> yuv2gbrp10be_full_X_4_512_c: 14401.6
> yuv2gbrp10be_full_X_4_512_sse2: 9133.1
> yuv2gbrp10be_full_X_4_512_sse4: 5126.1
> yuv2gbrp10be_full_X_4_512_avx2: 2837.6
> yuv2gbrp10le_full_X_4_512_c: 12718.1
> yuv2gbrp10le_full_X_4_512_sse2: 9106.1
> yuv2gbrp10le_full_X_4_512_sse4: 5120.1
> yuv2gbrp10le_full_X_4_512_avx2: 2826.1
> yuv2gbrap10be_full_X_4_512_c: 18535.6
> yuv2gbrap10be_full_X_4_512_sse2: 33617.6
> yuv2gbrap10be_full_X_4_512_sse4: 6264.1
> yuv2gbrap10be_full_X_4_512_avx2: 3422.1
> yuv2gbrap10le_full_X_4_512_c: 16724.1
> yuv2gbrap10le_full_X_4_512_sse2: 11787.1
> yuv2gbrap10le_full_X_4_512_sse4: 6282.1
> yuv2gbrap10le_full_X_4_512_avx2: 3441.6
> yuv2gbrp12be_full_X_4_512_c: 13723.6
> yuv2gbrp12be_full_X_4_512_sse2: 9128.1
> yuv2gbrp12be_full_X_4_512_sse4: 7997.6
> yuv2gbrp12be_full_X_4_512_avx2: 2844.1
> yuv2gbrp12le_full_X_4_512_c: 12257.1
> yuv2gbrp12le_full_X_4_512_sse2: 9107.6
> yuv2gbrp12le_full_X_4_512_sse4: 5142.6
> yuv2gbrp12le_full_X_4_512_avx2: 2837.6
> yuv2gbrap12be_full_X_4_512_c: 18511.1
> yuv2gbrap12be_full_X_4_512_sse2: 12156.6
> yuv2gbrap12be_full_X_4_512_sse4: 6251.1
> yuv2gbrap12be_full_X_4_512_avx2: 3444.6
> yuv2gbrap12le_full_X_4_512_c: 16687.1
> yuv2gbrap12le_full_X_4_512_sse2: 11785.1
> yuv2gbrap12le_full_X_4_512_sse4: 6243.6
> yuv2gbrap12le_full_X_4_512_avx2: 3446.1
> yuv2gbrp14be_full_X_4_512_c: 13690.6
> yuv2gbrp14be_full_X_4_512_sse2: 9120.6
> yuv2gbrp14be_full_X_4_512_sse4: 5138.1
> yuv2gbrp14be_full_X_4_512_avx2: 2843.1
> yuv2gbrp14le_full_X_4_512_c: 14995.6
> yuv2gbrp14le_full_X_4_512_sse2: 9119.1
> yuv2gbrp14le_full_X_4_512_sse4: 5126.1
> yuv2gbrp14le_full_X_4_512_avx2: 2843.1
> yuv2gbrp16be_full_X_4_512_c: 12367.1
> yuv2gbrp16be_full_X_4_512_sse2: 8233.6
> yuv2gbrp16be_full_X_4_512_sse4: 4820.1
> yuv2gbrp16be_full_X_4_512_avx2: 2666.6
> yuv2gbrp16le_full_X_4_512_c: 10904.1
> yuv2gbrp16le_full_X_4_512_sse2: 8214.1
> yuv2gbrp16le_full_X_4_512_sse4: 4824.1
> yuv2gbrp16le_full_X_4_512_avx2: 2629.1
> yuv2gbrap16be_full_X_4_512_c: 26569.6
> yuv2gbrap16be_full_X_4_512_sse2: 10884.1
> yuv2gbrap16be_full_X_4_512_sse4: 5488.1
> yuv2gbrap16be_full_X_4_512_avx2: 3272.1
> yuv2gbrap16le_full_X_4_512_c: 14010.1
> yuv2gbrap16le_full_X_4_512_sse2: 10562.1
> yuv2gbrap16le_full_X_4_512_sse4: 5463.6
> yuv2gbrap16le_full_X_4_512_avx2: 3255.1
> yuv2gbrpf32be_full_X_4_512_c: 14524.1
> yuv2gbrpf32be_full_X_4_512_sse2: 8552.6
> yuv2gbrpf32be_full_X_4_512_sse4: 4636.1
> yuv2gbrpf32be_full_X_4_512_avx2: 2474.6
> yuv2gbrpf32le_full_X_4_512_c: 13060.6
> yuv2gbrpf32le_full_X_4_512_sse2: 9682.6
> yuv2gbrpf32le_full_X_4_512_sse4: 4298.1
> yuv2gbrpf32le_full_X_4_512_avx2: 2453.1
> yuv2gbrapf32be_full_X_4_512_c: 18629.6
> yuv2gbrapf32be_full_X_4_512_sse2: 11363.1
> yuv2gbrapf32be_full_X_4_512_sse4: 15201.6
> yuv2gbrapf32be_full_X_4_512_avx2: 3727.1
> yuv2gbrapf32le_full_X_4_512_c: 16677.6
> yuv2gbrapf32le_full_X_4_512_sse2: 10221.6
> yuv2gbrapf32le_full_X_4_512_sse4: 5693.6
> yuv2gbrapf32le_full_X_4_512_avx2: 3656.6
> ---
> libswscale/x86/output.asm | 434 +++++++++++++++++++++++++++++++++++++-
> libswscale/x86/swscale.c | 98 +++++++++
> tests/checkasm/Makefile | 2 +-
> tests/checkasm/checkasm.c | 1 +
> tests/checkasm/checkasm.h | 1 +
> tests/checkasm/sw_gbrp.c | 198 +++++++++++++++++
> tests/fate/checkasm.mak | 1 +
> 7 files changed, 733 insertions(+), 2 deletions(-)
> create mode 100644 tests/checkasm/sw_gbrp.c
>
> diff --git a/libswscale/x86/output.asm b/libswscale/x86/output.asm
> index 52cf9f2c2e..e277a61449 100644
> --- a/libswscale/x86/output.asm
> +++ b/libswscale/x86/output.asm
> @@ -38,7 +38,49 @@ pw_32: times 8 dw 32
> pd_255: times 8 dd 255
> pw_512: times 8 dw 512
> pw_1024: times 8 dw 1024
> -
> +pd_65535_invf: times 8 dd 0x37800080 ;1.0/65535.0
> +pd_yuv2gbrp16_start: times 8 dd -0x40000000
> +pd_yuv2gbrp_y_start: times 8 dd (1 << 9)
> +pd_yuv2gbrp_uv_start: times 8 dd ((1 << 9) - (128 << 19))
> +pd_yuv2gbrp_a_start: times 8 dd (1 << 18)
> +pd_yuv2gbrp16_offset: times 8 dd 0x10000 ;(1 << 16)
> +pd_yuv2gbrp16_round13: times 8 dd 0x02000 ;(1 << 13)
> +pd_yuv2gbrp16_a_offset: times 8 dd 0x20002000
> +pd_yuv2gbrp16_upper30: times 8 dd 0x3FFFFFFF ;(1<<30) - 1
> +pd_yuv2gbrp16_upper27: times 8 dd 0x07FFFFFF ;(1<<27) - 1
> +pd_yuv2gbrp16_upperC: times 8 dd 0xC0000000
> +pb_pack_shuffle8: db 0, 4, 8, 12, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + 0, 4, 8, 12, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1
> +pb_pack_shuffle16le: db 0, 1, 4, 5, \
> + 8, 9, 12, 13, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + 0, 1, 4, 5, \
> + 8, 9, 12, 13
> +pb_pack_shuffle16be: db 1, 0, 5, 4, \
> + 9, 8, 13, 12, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + -1, -1, -1, -1, \
> + 1, 0, 5, 4, \
> + 9, 8, 13, 12
> +pb_shuffle32be: db 3, 2, 1, 0, \
> + 7, 6, 5, 4, \
> + 11, 10, 9, 8, \
> + 15, 14, 13, 12, \
> + 3, 2, 1, 0, \
> + 7, 6, 5, 4, \
> + 11, 10, 9, 8, \
> + 15, 14, 13, 12
> yuv2nv12_shuffle_mask: times 2 db 0, 4, 8, 12, \
> -1, -1, -1, -1, \
> -1, -1, -1, -1, \
> @@ -549,3 +591,393 @@ yuv2nv12cX_fn yuv2nv12
> yuv2nv12cX_fn yuv2nv21
> %endif
> %endif ; ARCH_X86_64
> +
>
> +;-----------------------------------------------------------------------------
> +; planar grb yuv2anyX functions
> +; void ff_yuv2<gbr_format>_full_X_<opt>(SwsContext *c, const int16_t
> *lumFilter,
> +; const int16_t **lumSrcx, int
> lumFilterSize,
> +; const int16_t *chrFilter, const
> int16_t **chrUSrcx,
> +; const int16_t **chrVSrcx, int
> chrFilterSize,
> +; const int16_t **alpSrcx, uint8_t
> **dest,
> +; int dstW, int y)
>
> +;-----------------------------------------------------------------------------
> +
> +%if ARCH_X86_64
> +struc SwsContext
> + .padding: resb 40292 ; offsetof(SwsContext,
> yuv2rgb_y_offset)
> + .yuv2rgb_y_offset: resd 1
> + .yuv2rgb_y_coeff: resd 1
> + .yuv2rgb_v2r_coeff: resd 1
> + .yuv2rgb_v2g_coeff: resd 1
> + .yuv2rgb_u2g_coeff: resd 1
> + .yuv2rgb_u2b_coeff: resd 1
> +endstruc
> +
> +%define R m0
> +%define G m1
> +%define B m2
> +%define A m3
> +
> +%define Y m4
> +%define U m5
> +%define V m6
> +
> +; Clip a signed integer to an unsigned power of two range.
> +; av_clip_uintp2
> +; 1 - dest
> +; 2 - bit position to clip at
> +%macro CLIPP2 2
> + ; (~a) >> 31 & ((1<<p) - 1);
> + pcmpeqb m4, m4
> + pxor m4, %1
> + psrad m4, 31
> + movu m5, [pd_yuv2gbrp16_upper%2]
> + pand m4, m5
> +
> + ; (a & ~((1<<p) - 1)) == 0
> + pandn m5, %1
> + pxor m6, m6
> + pcmpeqd m5, m6
> +%if cpuflag(avx2)
> + vpblendvb %1, m4, %1, m5
> +%else
> + pxor %1, m4
> + pand %1, m5
> + pxor %1, m4
> +%endif
> +%endmacro
> +
> +; 1 - dest
> +; 2 - source
> +%macro LOAD16 2
> + %if cpuflag(avx2)
> + movu xm%1, %2
> + vpmovsxwd m%1, xm%1
> + %elif cpuflag(sse4)
> + movsd m%1, %2
> + pmovsxwd m%1, m%1
> + %else
> + movsd m%1, %2
> + punpcklwd m%1, m%1
> + psrad m%1, 16 ; sign extend
> + %endif
> +%endmacro
> +
> +; 1 - dest
> +; 2 - source
> +; 3 - depth
> +%macro LOAD_PIXELS 3
> + mov ptrq, [%2 + jq*8]
> +%if %3 >= 16
> + movu m%1, [ptrq + xq*4]
> +%else
> + LOAD16 %1, [ptrq + xq*2]
> +%endif
> +%endmacro
> +
> +; 1 - dest
> +; 2 - source
> +%macro STORE8 2
> + mov ptrq, %1
> + %if mmsize > 16
> + pshufb m%2, [pb_pack_shuffle8]
> + vextractf128 xm4, m%2, 1
> + por xm%2, xm4
> + movq [ptrq + xq], xm%2
> + %else
> + %if cpuflag(sse4)
> + pshufb m%2, [pb_pack_shuffle8]
> + %else
> + psrldq m4, m%2, 3
> + por m%2, m4
> + psrldq m4, m%2, 6
> + por m%2, m4
> + %endif
> + movd [ptrq + xq], m%2
> + %endif
> +%endmacro
> +
> +; 1 - dest
> +; 2 - source
> +; 3 - is big endian
> +%macro STORE16 3
> + mov ptrq, %1
> + %if mmsize > 16
> + %if %3 ; bigendian
> + pshufb m%2, [pb_pack_shuffle16be]
> + %else
> + pshufb m%2, [pb_pack_shuffle16le]
> + %endif
> + vpermq m%2, m%2, (3 << 6 | 0 << 4 | 3 << 2 | 0 << 0)
> + movu [ptrq + xq*2], xm%2
> + %else
> + %if cpuflag(sse4) && %3 ; bigendian
> + pshufb m%2, [pb_pack_shuffle16be]
> + %elif cpuflag(sse4)
> + pshufb m%2, [pb_pack_shuffle16le]
> + %else
> + pshuflw m%2, m%2, (1 << 6 | 1 << 4 | 2 << 2 | 0 << 0)
> + pshufhw m%2, m%2, (1 << 6 | 1 << 4 | 2 << 2 | 0 << 0)
> + pshufd m%2, m%2, (3 << 6 | 3 << 4 | 2 << 2 | 0 << 0)
> + %if %3 ; bigendian
> + psrlw m4, m%2, 8
> + psllw m%2, 8
> + por m%2, m4
> + %endif
> + %endif
> + movq [ptrq + xq*2], m%2
> + %endif
> +%endmacro
> +
> +%macro SWAP32 1
> +%if mmsize > 16 || cpuflag(sse4)
> + pshufb m%1, [pb_shuffle32be]
> +%else
> + psrlw m4, m%1, 8
> + psllw m%1, 8
> + por m%1, m4
> + pshuflw m%1, m%1, (2 << 6 | 3 << 4 | 0 << 2 | 1 << 0)
> + pshufhw m%1, m%1, (2 << 6 | 3 << 4 | 0 << 2 | 1 << 0)
> +%endif
> +%endmacro
> +
> +; 1 - dest
> +; 2 - source
> +; 3 - depth
> +; 4 - is big endian
> +%macro STORE_PIXELS 4
> +%if %3 > 16
> + %if %4
> + SWAP32 %2
> + %endif
> + mov ptrq, %1
> + movu [ptrq + xq*4], m%2
> +%elif %3 > 8
> + STORE16 %1, %2, %4
> +%else
> + STORE8 %1, %2
> +%endif
> +%endmacro
> +
> +%macro PMULLO 3
> +%if cpuflag(sse4) || mmsize > 16
> + pmulld %1, %2, %3
> +%else
> + %ifidni %1, %2
> + %else
> + mova %1, %2
> + %endif
> + pshufd m7, %1, (2 << 6 | 3 << 4 | 0 << 2 | 1 << 0) ; 0xb1
> + pshufd m8, %3, (2 << 6 | 3 << 4 | 0 << 2 | 1 << 0) ; 0xb1
> + pmuludq m7, m8
> + pshufd m7, m7, (3 << 6 | 1 << 4 | 2 << 2 | 0 << 0) ; 0xd8
> + pmuludq %1, %3
> + pshufd %1, %1, (3 << 6 | 1 << 4 | 2 << 2 | 0 << 0) ; 0xd8
> + punpckldq %1, m7
> +%endif
> +%endmacro
> +
> +; 1 - name
> +; 2 - depth
> +; 3 - has alpha
> +; 3 - is big endian
> +; 5 - is float
> +%macro yuv2gbrp_fn 5
> +%define DEPTH %2
> +%define HAS_ALPHA %3
> +%define IS_BE %4
> +%define FLOAT %5
> +%define SH (22 + 8 - DEPTH)
> +
> +%if DEPTH >= 16
> + %define RGB_SHIFT 14
> + %define A_SHIFT 14
> +%elif 22 != SH
> + %define RGB_SHIFT SH
> + %define A_SHIFT (SH-3)
> +%else
> + %define RGB_SHIFT 22
> + %define A_SHIFT 19
> +%endif
> +
> +%if DEPTH >= 16
> + %define YUV_SHIFT 14
> + %define Y_START m9
> + %define Y_ROUND [pd_yuv2gbrp16_round13]
> + %define UV_START m9
> + %define A_START m9
> + %define A_CLIP2P 30
> +%else
> + %define YUV_SHIFT 10
> + %define Y_START [pd_yuv2gbrp_y_start]
> + %define Y_ROUND m9
> + %define UV_START [pd_yuv2gbrp_uv_start]
> + %define A_START [pd_yuv2gbrp_a_start]
> + %define A_CLIP2P 27
> +%endif
> +
> +cglobal yuv2%1_full_X, 12, 14, 16, ptr, lumFilter, lumSrcx,
> lumFilterSize, chrFilter, chrUSrcx, chrVSrcx, chrFilterSize, alpSrcx, dest,
> dstW, y, x, j
> + VBROADCASTSS m10, dword [ptrq + SwsContext.yuv2rgb_y_offset]
> + VBROADCASTSS m11, dword [ptrq + SwsContext.yuv2rgb_y_coeff]
> + VBROADCASTSS m12, dword [ptrq + SwsContext.yuv2rgb_v2r_coeff]
> + VBROADCASTSS m13, dword [ptrq + SwsContext.yuv2rgb_v2g_coeff]
> + VBROADCASTSS m14, dword [ptrq + SwsContext.yuv2rgb_u2g_coeff]
> + VBROADCASTSS m15, dword [ptrq + SwsContext.yuv2rgb_u2b_coeff]
> +
> +%if DEPTH >= 16
> + movu m9, [pd_yuv2gbrp16_start]
> +%else
> + mov xq, (1 << (SH-1))
> + movq xm9, xq
> + VBROADCASTSS m9, xm9
> +%endif
> + xor xq, xq
> +
> + %%loop_x:
> + movu Y, Y_START
> + movu U, UV_START
> + movu V, UV_START
> +
> + xor jq, jq
> + %%loop_luma:
> + movsx ptrd, word [lumFilterq + jq*2]
> + movd xm0, ptrd
> + VBROADCASTSS m0, xm0
> + LOAD_PIXELS 1, lumSrcxq, DEPTH
> + PMULLO m1, m1, m0
> + paddd Y, m1
> + inc jd
> + cmp jd, lumFilterSized
> + jl %%loop_luma
> +
> +%if HAS_ALPHA
> + cmp alpSrcxq, 0
> + je %%skip_alpha_load
> + xor jq, jq
> + movu A, A_START
> + %%loop_alpha:
> + movsx ptrd, word [lumFilterq + jq*2]
> + movd xm0, ptrd
> + VBROADCASTSS m0, xm0
> + LOAD_PIXELS 1, alpSrcxq, DEPTH
> + PMULLO m1, m1, m0
> + paddd A, m1
> + inc jd
> + cmp jd, lumFilterSized
> + jl %%loop_alpha
> +%if DEPTH >= 16
> + psrad A, 1
> + paddd A, [pd_yuv2gbrp16_a_offset]
> +%endif
> + %%skip_alpha_load:
> +%endif
> + xor jq, jq
> + %%loop_chr:
> + movsx ptrd, word [chrFilterq + jq*2]
> + movd xm0, ptrd
> + VBROADCASTSS m0, xm0
> + LOAD_PIXELS 1, chrUSrcxq, DEPTH
> + LOAD_PIXELS 2, chrVSrcxq, DEPTH
> + PMULLO m1, m1, m0
> + PMULLO m2, m2, m0
> + paddd U, m1
> + paddd V, m2
> + inc jd
> + cmp jd, chrFilterSized
> + jl %%loop_chr
> +
> + psrad Y, YUV_SHIFT
> +%if DEPTH >= 16
> + paddd Y, [pd_yuv2gbrp16_offset]
> +%endif
> + psrad U, YUV_SHIFT
> + psrad V, YUV_SHIFT
> +
> + psubd Y, m10 ; yuv2rgb_y_offset
> + PMULLO Y, Y, m11 ; yuv2rgb_y_coeff
> + paddd Y, Y_ROUND
> +
> + PMULLO R, V, m12 ; yuv2rgb_v2r_coeff
> + PMULLO B, U, m15 ; yuv2rgb_u2b_coeff
> +
> + PMULLO U, U, m14 ; yuv2rgb_u2g_coeff
> + PMULLO V, V, m13 ; yuv2rgb_v2g_coeff
> + paddd G, U, V
> + paddd R, Y
> + paddd G, Y
> + paddd B, Y
> +
> + CLIPP2 R, 30
> + CLIPP2 G, 30
> + CLIPP2 B, 30
> +
> + psrad R, RGB_SHIFT
> + psrad G, RGB_SHIFT
> + psrad B, RGB_SHIFT
> +
> +%if FLOAT
> + cvtdq2ps R, R
> + cvtdq2ps G, G
> + cvtdq2ps B, B
> + mulps R, [pd_65535_invf]
> + mulps G, [pd_65535_invf]
> + mulps B, [pd_65535_invf]
> +%endif
> + STORE_PIXELS [destq + 0], 1, DEPTH, IS_BE ; G
> + STORE_PIXELS [destq + 8], 2, DEPTH, IS_BE ; B
> + STORE_PIXELS [destq + 16], 0, DEPTH, IS_BE ; R
> +
> +%if HAS_ALPHA
> + cmp alpSrcxq, 0
> + je %%skip_alpha_store
> + CLIPP2 A, A_CLIP2P
> + psrad A, A_SHIFT
> +%if FLOAT
> + cvtdq2ps A, A
> + mulps A, [pd_65535_invf]
> +%endif
> + STORE_PIXELS [destq + 24], 3, DEPTH, IS_BE
> + %%skip_alpha_store:
> +%endif
> + add xq, mmsize/4
> + cmp xd, dstWd
> + jl %%loop_x
> +
> + RET
> +%endmacro
> +
> +%macro yuv2gbrp_fn_decl 2
> +INIT_%1 %2
> +yuv2gbrp_fn gbrp, 8, 0, 0, 0
> +yuv2gbrp_fn gbrap, 8, 1, 0, 0
> +yuv2gbrp_fn gbrp9le, 9, 0, 0, 0
> +yuv2gbrp_fn gbrp10le, 10, 0, 0, 0
> +yuv2gbrp_fn gbrap10le, 10, 1, 0, 0
> +yuv2gbrp_fn gbrp12le, 12, 0, 0, 0
> +yuv2gbrp_fn gbrap12le, 12, 1, 0, 0
> +yuv2gbrp_fn gbrp14le, 14, 0, 0, 0
> +yuv2gbrp_fn gbrp16le, 16, 0, 0, 0
> +yuv2gbrp_fn gbrap16le, 16, 1, 0, 0
> +yuv2gbrp_fn gbrpf32le, 32, 0, 0, 1
> +yuv2gbrp_fn gbrapf32le, 32, 1, 0, 1
> +
> +yuv2gbrp_fn gbrp9be, 9, 0, 1, 0
> +yuv2gbrp_fn gbrp10be, 10, 0, 1, 0
> +yuv2gbrp_fn gbrap10be, 10, 1, 1, 0
> +yuv2gbrp_fn gbrp12be, 12, 0, 1, 0
> +yuv2gbrp_fn gbrap12be, 12, 1, 1, 0
> +yuv2gbrp_fn gbrp14be, 14, 0, 1, 0
> +yuv2gbrp_fn gbrp16be, 16, 0, 1, 0
> +yuv2gbrp_fn gbrap16be, 16, 1, 1, 0
> +yuv2gbrp_fn gbrpf32be, 32, 0, 1, 1
> +yuv2gbrp_fn gbrapf32be, 32, 1, 1, 1
> +%endmacro
> +
> +yuv2gbrp_fn_decl XMM, sse2
> +yuv2gbrp_fn_decl XMM, sse4
> +
> +%if HAVE_AVX2_EXTERNAL
> +yuv2gbrp_fn_decl YMM, avx2
> +%endif
> +
> +%endif ; ARCH_X86_64
> \ No newline at end of file
> diff --git a/libswscale/x86/swscale.c b/libswscale/x86/swscale.c
> index 0848a31461..e8bc15b6ee 100644
> --- a/libswscale/x86/swscale.c
> +++ b/libswscale/x86/swscale.c
> @@ -350,6 +350,43 @@ void ff_yuv2 ## fmt ## cX_ ## opt(enum AVPixelFormat
> format, const uint8_t *dith
>
> YUV2NV_DECL(nv12, avx2);
> YUV2NV_DECL(nv21, avx2);
> +
> +#define YUV2GBRP_FN_DECL(fmt, opt)
> \
> +void ff_yuv2##fmt##_full_X_ ##opt(SwsContext *c, const int16_t
> *lumFilter, \
> + const int16_t **lumSrcx, int
> lumFilterSize, \
> + const int16_t *chrFilter, const int16_t
> **chrUSrcx, \
> + const int16_t **chrVSrcx, int
> chrFilterSize, \
> + const int16_t **alpSrcx, uint8_t
> **dest, \
> + int dstW, int y)
> +
> +#define YUV2GBRP_DECL(opt) \
> +YUV2GBRP_FN_DECL(gbrp, opt); \
> +YUV2GBRP_FN_DECL(gbrap, opt); \
> +YUV2GBRP_FN_DECL(gbrp9le, opt); \
> +YUV2GBRP_FN_DECL(gbrp10le, opt); \
> +YUV2GBRP_FN_DECL(gbrap10le, opt); \
> +YUV2GBRP_FN_DECL(gbrp12le, opt); \
> +YUV2GBRP_FN_DECL(gbrap12le, opt); \
> +YUV2GBRP_FN_DECL(gbrp14le, opt); \
> +YUV2GBRP_FN_DECL(gbrp16le, opt); \
> +YUV2GBRP_FN_DECL(gbrap16le, opt); \
> +YUV2GBRP_FN_DECL(gbrpf32le, opt); \
> +YUV2GBRP_FN_DECL(gbrapf32le, opt); \
> +YUV2GBRP_FN_DECL(gbrp9be, opt); \
> +YUV2GBRP_FN_DECL(gbrp10be, opt); \
> +YUV2GBRP_FN_DECL(gbrap10be, opt); \
> +YUV2GBRP_FN_DECL(gbrp12be, opt); \
> +YUV2GBRP_FN_DECL(gbrap12be, opt); \
> +YUV2GBRP_FN_DECL(gbrp14be, opt); \
> +YUV2GBRP_FN_DECL(gbrp16be, opt); \
> +YUV2GBRP_FN_DECL(gbrap16be, opt); \
> +YUV2GBRP_FN_DECL(gbrpf32be, opt); \
> +YUV2GBRP_FN_DECL(gbrapf32be, opt);
> +
> +YUV2GBRP_DECL(sse2);
> +YUV2GBRP_DECL(sse4);
> +YUV2GBRP_DECL(avx2);
> +
> #endif
>
> av_cold void ff_sws_init_swscale_x86(SwsContext *c)
> @@ -582,5 +619,66 @@ switch(c->dstBpc){ \
> break;
> }
> }
> +
> + if(c->flags & SWS_FULL_CHR_H_INT) {
> +
> + /* yuv2gbrp uses the SwsContext for yuv coefficients
> + if struct offsets change the asm needs to be updated too */
> + av_assert0(offsetof(SwsContext, yuv2rgb_y_offset) == 40292);
> +
> +#define YUV2ANYX_FUNC_CASE(fmt, name, opt) \
> + case fmt: \
> + c->yuv2anyX = ff_yuv2##name##_full_X_##opt; \
> + break;
> +
> +#define YUV2ANYX_GBRAP_CASES(opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP, gbrp, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP, gbrap, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP9LE, gbrp9le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP10LE, gbrp10le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP10LE, gbrap10le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP12LE, gbrp12le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP12LE, gbrap12le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP14LE, gbrp14le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP16LE, gbrp16le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP16LE, gbrap16le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRPF32LE, gbrpf32le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAPF32LE, gbrapf32le, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP9BE, gbrp9be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP10BE, gbrp10be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP10BE, gbrap10be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP12BE, gbrp12be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP12BE, gbrap12be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP14BE, gbrp14be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRP16BE, gbrp16be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAP16BE, gbrap16be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRPF32BE, gbrpf32be, opt) \
> + YUV2ANYX_FUNC_CASE(AV_PIX_FMT_GBRAPF32BE, gbrapf32be, opt)
> +
> + if (EXTERNAL_SSE2(cpu_flags)) {
> + switch (c->dstFormat) {
> + YUV2ANYX_GBRAP_CASES(sse2)
> + default:
> + break;
> + }
> + }
> +
> + if (EXTERNAL_SSE4(cpu_flags)) {
> + switch (c->dstFormat) {
> + YUV2ANYX_GBRAP_CASES(sse4)
> + default:
> + break;
> + }
> + }
> +
> + if (EXTERNAL_AVX2_FAST(cpu_flags)) {
> + switch (c->dstFormat) {
> + YUV2ANYX_GBRAP_CASES(avx2)
> + default:
> + break;
> + }
> + }
> + }
> +
> #endif
> }
> diff --git a/tests/checkasm/Makefile b/tests/checkasm/Makefile
> index 4ef5fa87da..f768b1144e 100644
> --- a/tests/checkasm/Makefile
> +++ b/tests/checkasm/Makefile
> @@ -45,7 +45,7 @@ AVFILTEROBJS-$(CONFIG_NLMEANS_FILTER) += vf_nlmeans.o
> CHECKASMOBJS-$(CONFIG_AVFILTER) += $(AVFILTEROBJS-yes)
>
> # swscale tests
> -SWSCALEOBJS += sw_rgb.o sw_scale.o
> +SWSCALEOBJS += sw_gbrp.o sw_rgb.o sw_scale.o
>
> CHECKASMOBJS-$(CONFIG_SWSCALE) += $(SWSCALEOBJS)
>
> diff --git a/tests/checkasm/checkasm.c b/tests/checkasm/checkasm.c
> index b1353f7cbe..5fa8ac7bcc 100644
> --- a/tests/checkasm/checkasm.c
> +++ b/tests/checkasm/checkasm.c
> @@ -183,6 +183,7 @@ static const struct {
> #endif
> #endif
> #if CONFIG_SWSCALE
> + { "sw_gbrp", checkasm_check_sw_gbrp },
> { "sw_rgb", checkasm_check_sw_rgb },
> { "sw_scale", checkasm_check_sw_scale },
> #endif
> diff --git a/tests/checkasm/checkasm.h b/tests/checkasm/checkasm.h
> index 68b0697d3e..c3192d8c23 100644
> --- a/tests/checkasm/checkasm.h
> +++ b/tests/checkasm/checkasm.h
> @@ -72,6 +72,7 @@ void checkasm_check_opusdsp(void);
> void checkasm_check_pixblockdsp(void);
> void checkasm_check_sbrdsp(void);
> void checkasm_check_synth_filter(void);
> +void checkasm_check_sw_gbrp(void);
> void checkasm_check_sw_rgb(void);
> void checkasm_check_sw_scale(void);
> void checkasm_check_utvideodsp(void);
> diff --git a/tests/checkasm/sw_gbrp.c b/tests/checkasm/sw_gbrp.c
> new file mode 100644
> index 0000000000..61b2b9d6a1
> --- /dev/null
> +++ b/tests/checkasm/sw_gbrp.c
> @@ -0,0 +1,198 @@
> +/*
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * FFmpeg is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License along
> + * with FFmpeg; if not, write to the Free Software Foundation, Inc.,
> + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
> + */
> +
> +#include <string.h>
> +
> +#include "libavutil/common.h"
> +#include "libavutil/intreadwrite.h"
> +#include "libavutil/mem_internal.h"
> +#include "libavutil/pixdesc.h"
> +
> +#include "libswscale/swscale.h"
> +#include "libswscale/swscale_internal.h"
> +
> +#include "checkasm.h"
> +
> +#define randomize_buffers(buf, size) \
> + do { \
> + int j; \
> + for (j = 0; j < size; j+=4) \
> + AV_WN32(buf + j, rnd()); \
> + } while (0)
> +
> +static const int planar_fmts[] = {
> + AV_PIX_FMT_GBRP,
> + AV_PIX_FMT_GBRP9BE,
> + AV_PIX_FMT_GBRP9LE,
> + AV_PIX_FMT_GBRP10BE,
> + AV_PIX_FMT_GBRP10LE,
> + AV_PIX_FMT_GBRP12BE,
> + AV_PIX_FMT_GBRP12LE,
> + AV_PIX_FMT_GBRP14BE,
> + AV_PIX_FMT_GBRP14LE,
> + AV_PIX_FMT_GBRAP,
> + AV_PIX_FMT_GBRAP10BE,
> + AV_PIX_FMT_GBRAP10LE,
> + AV_PIX_FMT_GBRAP12BE,
> + AV_PIX_FMT_GBRAP12LE,
> + AV_PIX_FMT_GBRP16BE,
> + AV_PIX_FMT_GBRP16LE,
> + AV_PIX_FMT_GBRAP16BE,
> + AV_PIX_FMT_GBRAP16LE,
> + AV_PIX_FMT_GBRPF32BE,
> + AV_PIX_FMT_GBRPF32LE,
> + AV_PIX_FMT_GBRAPF32BE,
> + AV_PIX_FMT_GBRAPF32LE
> +};
> +
> +static void check_output_yuv2gbrp(void)
> +{
> + struct SwsContext *ctx;
> + const AVPixFmtDescriptor *desc;
> + int fmi, fsi, isi, i;
> + int dstW, byte_size, luma_filter_size, chr_filter_size;
> +#define LARGEST_FILTER 16
> +#define FILTER_SIZES 4
> + static const int filter_sizes[] = {1, 4, 8, 16};
> +#define LARGEST_INPUT_SIZE 512
> +#define INPUT_SIZES 6
> + static const int input_sizes[] = {8, 24, 128, 144, 256, 512};
> + uint8_t *dst0[4];
> + uint8_t *dst1[4];
> +
> + declare_func(void, void *c, const int16_t *lumFilter,
> + const int16_t **lumSrcx, int lumFilterSize,
> + const int16_t *chrFilter, const int16_t **chrUSrcx,
> + const int16_t **chrVSrcx, int chrFilterSize,
> + const int16_t **alpSrcx, uint8_t **dest,
> + int dstW, int y);
> +
> + const int16_t *luma[LARGEST_FILTER];
> + const int16_t *chru[LARGEST_FILTER];
> + const int16_t *chrv[LARGEST_FILTER];
> + const int16_t *alpha[LARGEST_FILTER];
> +
> + LOCAL_ALIGNED_8(int16_t, luma_filter, [LARGEST_FILTER]);
> + LOCAL_ALIGNED_8(int16_t, chr_filter, [LARGEST_FILTER]);
> +
> + LOCAL_ALIGNED_8(int32_t, src_y, [LARGEST_FILTER *
> LARGEST_INPUT_SIZE]);
> + LOCAL_ALIGNED_8(int32_t, src_u, [LARGEST_FILTER *
> LARGEST_INPUT_SIZE]);
> + LOCAL_ALIGNED_8(int32_t, src_v, [LARGEST_FILTER *
> LARGEST_INPUT_SIZE]);
> + LOCAL_ALIGNED_8(int32_t, src_a, [LARGEST_FILTER *
> LARGEST_INPUT_SIZE]);
> +
> + LOCAL_ALIGNED_8(uint8_t, dst0_r, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst0_g, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst0_b, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst0_a, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> +
> + LOCAL_ALIGNED_8(uint8_t, dst1_r, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst1_g, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst1_b, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> + LOCAL_ALIGNED_8(uint8_t, dst1_a, [LARGEST_INPUT_SIZE *
> sizeof(int32_t)]);
> +
> + randomize_buffers((uint8_t*)src_y, LARGEST_FILTER *
> LARGEST_INPUT_SIZE * sizeof(int32_t));
> + randomize_buffers((uint8_t*)src_u, LARGEST_FILTER *
> LARGEST_INPUT_SIZE * sizeof(int32_t));
> + randomize_buffers((uint8_t*)src_v, LARGEST_FILTER *
> LARGEST_INPUT_SIZE * sizeof(int32_t));
> + randomize_buffers((uint8_t*)src_a, LARGEST_FILTER *
> LARGEST_INPUT_SIZE * sizeof(int32_t));
> + randomize_buffers((uint8_t*)luma_filter, LARGEST_FILTER *
> sizeof(int16_t));
> + randomize_buffers((uint8_t*)chr_filter, LARGEST_FILTER *
> sizeof(int16_t));
> +
> + dst0[0] = (uint8_t*)dst0_g;
> + dst0[1] = (uint8_t*)dst0_b;
> + dst0[2] = (uint8_t*)dst0_r;
> + dst0[3] = (uint8_t*)dst0_a;
> +
> + dst1[0] = (uint8_t*)dst1_g;
> + dst1[1] = (uint8_t*)dst1_b;
> + dst1[2] = (uint8_t*)dst1_r;
> + dst1[3] = (uint8_t*)dst1_a;
> +
> + for (i = 0; i < LARGEST_FILTER; i++) {
> + luma[i] = (int16_t *)(src_y + i*LARGEST_INPUT_SIZE);
> + chru[i] = (int16_t *)(src_u + i*LARGEST_INPUT_SIZE);
> + chrv[i] = (int16_t *)(src_v + i*LARGEST_INPUT_SIZE);
> + alpha[i] = (int16_t *)(src_a + i*LARGEST_INPUT_SIZE);
> + }
> +
> + ctx = sws_alloc_context();
> + if (sws_init_context(ctx, NULL, NULL) < 0)
> + fail();
> +
> + ctx->flags |= SWS_FULL_CHR_H_INT;
> + ctx->yuv2rgb_y_offset = rnd();
> + ctx->yuv2rgb_y_coeff = rnd();
> + ctx->yuv2rgb_v2r_coeff = rnd();
> + ctx->yuv2rgb_v2g_coeff = rnd();
> + ctx->yuv2rgb_u2g_coeff = rnd();
> + ctx->yuv2rgb_u2b_coeff = rnd();
> +
> + for (fmi = 0; fmi < FF_ARRAY_ELEMS(planar_fmts); fmi++) {
> + for (fsi = 0; fsi < FILTER_SIZES; fsi++) {
> + for (isi = 0; isi < INPUT_SIZES; isi++ ) {
> + desc = av_pix_fmt_desc_get(planar_fmts[fmi]);
> + ctx->dstFormat = planar_fmts[fmi];
> +
> + dstW = input_sizes[isi];
> + luma_filter_size = filter_sizes[fsi];
> + chr_filter_size = filter_sizes[fsi];
> +
> + if (desc->comp[0].depth > 16) {
> + byte_size = 4;
> + } else if (desc->comp[0].depth > 8) {
> + byte_size = 2;
> + } else {
> + byte_size = 1;
> + }
> +
> + ff_sws_init_scale(ctx);
> + if (check_func(ctx->yuv2anyX, "yuv2%s_full_X_%d_%d",
> desc->name, luma_filter_size, dstW)) {
> + for (i = 0; i < 4; i ++) {
> + memset(dst0[i], 0xFF, LARGEST_INPUT_SIZE *
> sizeof(int32_t));
> + memset(dst1[i], 0xFF, LARGEST_INPUT_SIZE *
> sizeof(int32_t));
> + }
> +
> + call_ref(ctx, luma_filter, luma, luma_filter_size,
> + chr_filter, chru, chrv, chr_filter_size,
> + alpha, dst0, dstW, 0);
> + call_new(ctx, luma_filter, luma, luma_filter_size,
> + chr_filter, chru, chrv, chr_filter_size,
> + alpha, dst1, dstW, 0);
> +
> + if (memcmp(dst0[0], dst1[0], dstW * byte_size) ||
> + memcmp(dst0[1], dst1[1], dstW * byte_size) ||
> + memcmp(dst0[2], dst1[2], dstW * byte_size) ||
> + memcmp(dst0[3], dst1[3], dstW * byte_size) )
> + fail();
> +
> + bench_new(ctx, luma_filter, luma, luma_filter_size,
> + chr_filter, chru, chrv, chr_filter_size,
> + alpha, dst1, dstW, 0);
> + }
> + }
> + }
> + }
> + sws_freeContext(ctx);
> +}
> +
> +
> +void checkasm_check_sw_gbrp(void)
> +{
> + check_output_yuv2gbrp();
> + report("output_yuv2gbrp");
> +}
> diff --git a/tests/fate/checkasm.mak b/tests/fate/checkasm.mak
> index 6e7edbe655..6db8f09d12 100644
> --- a/tests/fate/checkasm.mak
> +++ b/tests/fate/checkasm.mak
> @@ -26,6 +26,7 @@ FATE_CHECKASM = fate-checkasm-aacpsdsp
> \
> fate-checkasm-pixblockdsp \
> fate-checkasm-sbrdsp \
> fate-checkasm-synth_filter \
> + fate-checkasm-sw_gbrp \
> fate-checkasm-sw_rgb \
> fate-checkasm-sw_scale \
> fate-checkasm-utvideodsp \
> --
> 2.31.1.windows.1
>
>
ping
More information about the ffmpeg-devel
mailing list