Skip to content

F
ffmpeg.wasm-core
Commit 6405ca7d authored by Diego Biurrun's avatar Diego Biurrun
Browse files
Options

x86: h264_idct: Update comments to match 8/10-bit depth optimization split

parent 3e2fa991
Hide whitespace changes
Inline Side-by-side
Showing with 42 additions and 32 deletions
libavcodec/x86/h264_idct.asm
View file @ 6405ca7d
...@@ -80,7 +80,7 @@ SECTION .text ...@@ -80,7 +80,7 @@ SECTION .text
%endmacro %endmacro
INIT_MMX mmx INIT_MMX mmx
; ff_h264_idct_add_mmx(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct_add_8_mmx(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct_add_8, 3, 3, 0 cglobal h264_idct_add_8, 3, 3, 0
IDCT4_ADD r0, r1, r2 IDCT4_ADD r0, r1, r2
RET RET
...@@ -202,7 +202,7 @@ cglobal h264_idct_add_8, 3, 3, 0 ...@@ -202,7 +202,7 @@ cglobal h264_idct_add_8, 3, 3, 0
%endmacro %endmacro
INIT_MMX mmx INIT_MMX mmx
; ff_h264_idct8_add_mmx(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct8_add_8_mmx(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct8_add_8, 3, 4, 0 cglobal h264_idct8_add_8, 3, 4, 0
%assign pad 128+4-(stack_offset&7) %assign pad 128+4-(stack_offset&7)
SUB rsp, pad SUB rsp, pad
...@@ -270,7 +270,7 @@ cglobal h264_idct8_add_8, 3, 4, 0 ...@@ -270,7 +270,7 @@ cglobal h264_idct8_add_8, 3, 4, 0
%endmacro %endmacro
INIT_XMM sse2 INIT_XMM sse2
; ff_h264_idct8_add_sse2(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct8_add_8_sse2(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct8_add_8, 3, 4, 10 cglobal h264_idct8_add_8, 3, 4, 10
IDCT8_ADD_SSE r0, r1, r2, r3 IDCT8_ADD_SSE r0, r1, r2, r3
RET RET
...@@ -307,7 +307,7 @@ cglobal h264_idct8_add_8, 3, 4, 10 ...@@ -307,7 +307,7 @@ cglobal h264_idct8_add_8, 3, 4, 10
%endmacro %endmacro
INIT_MMX mmxext INIT_MMX mmxext
; ff_h264_idct_dc_add_mmxext(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct_dc_add_8_mmxext(uint8_t *dst, int16_t *block, int stride)
%if ARCH_X86_64 %if ARCH_X86_64
cglobal h264_idct_dc_add_8, 3, 4, 0 cglobal h264_idct_dc_add_8, 3, 4, 0
movsx r3, word [r1] movsx r3, word [r1]
...@@ -316,7 +316,7 @@ cglobal h264_idct_dc_add_8, 3, 4, 0 ...@@ -316,7 +316,7 @@ cglobal h264_idct_dc_add_8, 3, 4, 0
DC_ADD_MMXEXT_OP movh, r0, r2, r3 DC_ADD_MMXEXT_OP movh, r0, r2, r3
RET RET
; ff_h264_idct8_dc_add_mmxext(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct8_dc_add_8_mmxext(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct8_dc_add_8, 3, 4, 0 cglobal h264_idct8_dc_add_8, 3, 4, 0
movsx r3, word [r1] movsx r3, word [r1]
mov dword [r1], 0 mov dword [r1], 0
...@@ -326,6 +326,7 @@ cglobal h264_idct8_dc_add_8, 3, 4, 0 ...@@ -326,6 +326,7 @@ cglobal h264_idct8_dc_add_8, 3, 4, 0
DC_ADD_MMXEXT_OP mova, r0, r2, r3 DC_ADD_MMXEXT_OP mova, r0, r2, r3
RET RET
%else %else
; ff_h264_idct_dc_add_8_mmxext(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct_dc_add_8, 2, 3, 0 cglobal h264_idct_dc_add_8, 2, 3, 0
movsx r2, word [r1] movsx r2, word [r1]
mov dword [r1], 0 mov dword [r1], 0
...@@ -334,7 +335,7 @@ cglobal h264_idct_dc_add_8, 2, 3, 0 ...@@ -334,7 +335,7 @@ cglobal h264_idct_dc_add_8, 2, 3, 0
DC_ADD_MMXEXT_OP movh, r0, r1, r2 DC_ADD_MMXEXT_OP movh, r0, r1, r2
RET RET
; ff_h264_idct8_dc_add_mmxext(uint8_t *dst, int16_t *block, int stride) ; ff_h264_idct8_dc_add_8_mmxext(uint8_t *dst, int16_t *block, int stride)
cglobal h264_idct8_dc_add_8, 2, 3, 0 cglobal h264_idct8_dc_add_8, 2, 3, 0
movsx r2, word [r1] movsx r2, word [r1]
mov dword [r1], 0 mov dword [r1], 0
...@@ -347,8 +348,9 @@ cglobal h264_idct8_dc_add_8, 2, 3, 0 ...@@ -347,8 +348,9 @@ cglobal h264_idct8_dc_add_8, 2, 3, 0
%endif %endif
INIT_MMX mmx INIT_MMX mmx
; ff_h264_idct_add16_mmx(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16_8_mmx(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg cglobal h264_idct_add16_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg
xor r5, r5 xor r5, r5
%ifdef PIC %ifdef PIC
...@@ -369,8 +371,9 @@ cglobal h264_idct_add16_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, ...@@ -369,8 +371,9 @@ cglobal h264_idct_add16_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride,
jl .nextblock jl .nextblock
REP_RET REP_RET
; ff_h264_idct8_add4_mmx(uint8_t *dst, const int *block_offset, ; ff_h264_idct8_add4_8_mmx(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct8_add4_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg cglobal h264_idct8_add4_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg
%assign pad 128+4-(stack_offset&7) %assign pad 128+4-(stack_offset&7)
SUB rsp, pad SUB rsp, pad
...@@ -402,8 +405,9 @@ cglobal h264_idct8_add4_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, ...@@ -402,8 +405,9 @@ cglobal h264_idct8_add4_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride,
RET RET
INIT_MMX mmxext INIT_MMX mmxext
; ff_h264_idct_add16_mmxext(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16_8_mmxext(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
xor r5, r5 xor r5, r5
%ifdef PIC %ifdef PIC
...@@ -448,8 +452,9 @@ cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride ...@@ -448,8 +452,9 @@ cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride
REP_RET REP_RET
INIT_MMX mmx INIT_MMX mmx
; ff_h264_idct_add16intra_mmx(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16intra_8_mmx(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16intra_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg cglobal h264_idct_add16intra_8, 5, 7 + npicregs, 0, dst, block_offset, block, stride, nnzc, cntr, coeff, picreg
xor r5, r5 xor r5, r5
%ifdef PIC %ifdef PIC
...@@ -472,9 +477,9 @@ cglobal h264_idct_add16intra_8, 5, 7 + npicregs, 0, dst, block_offset, block, st ...@@ -472,9 +477,9 @@ cglobal h264_idct_add16intra_8, 5, 7 + npicregs, 0, dst, block_offset, block, st
REP_RET REP_RET
INIT_MMX mmxext INIT_MMX mmxext
; ff_h264_idct_add16intra_mmxext(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16intra_8_mmxext(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, ; int16_t *block, int stride,
; const uint8_t nnzc[6*8]) ; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
xor r5, r5 xor r5, r5
%ifdef PIC %ifdef PIC
...@@ -516,9 +521,9 @@ cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1, block_offset, block, s ...@@ -516,9 +521,9 @@ cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1, block_offset, block, s
jl .nextblock jl .nextblock
REP_RET REP_RET
; ff_h264_idct8_add4_mmxext(uint8_t *dst, const int *block_offset, ; ff_h264_idct8_add4_8_mmxext(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, ; int16_t *block, int stride,
; const uint8_t nnzc[6*8]) ; const uint8_t nnzc[6 * 8])
cglobal h264_idct8_add4_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct8_add4_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
%assign pad 128+4-(stack_offset&7) %assign pad 128+4-(stack_offset&7)
SUB rsp, pad SUB rsp, pad
...@@ -578,8 +583,9 @@ cglobal h264_idct8_add4_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride ...@@ -578,8 +583,9 @@ cglobal h264_idct8_add4_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride
RET RET
INIT_XMM sse2 INIT_XMM sse2
; ff_h264_idct8_add4_sse2(uint8_t *dst, const int *block_offset, ; ff_h264_idct8_add4_8_sse2(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct8_add4_8, 5, 8 + npicregs, 10, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct8_add4_8, 5, 8 + npicregs, 10, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
xor r5, r5 xor r5, r5
%ifdef PIC %ifdef PIC
...@@ -654,8 +660,8 @@ h264_idct_add8_mmx_plane: ...@@ -654,8 +660,8 @@ h264_idct_add8_mmx_plane:
jnz .nextblock jnz .nextblock
rep ret rep ret
; ff_h264_idct_add8_mmx(uint8_t **dest, const int *block_offset, ; ff_h264_idct_add8_8_mmx(uint8_t **dest, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride, const uint8_t nnzc[6 * 8])
cglobal h264_idct_add8_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct_add8_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
mov r5, 16 mov r5, 16
add r2, 512 add r2, 512
...@@ -719,8 +725,9 @@ h264_idct_add8_mmxext_plane: ...@@ -719,8 +725,9 @@ h264_idct_add8_mmxext_plane:
rep ret rep ret
INIT_MMX mmxext INIT_MMX mmxext
; ff_h264_idct_add8_mmxext(uint8_t **dest, const int *block_offset, ; ff_h264_idct_add8_8_mmxext(uint8_t **dest, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add8_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg cglobal h264_idct_add8_8, 5, 8 + npicregs, 0, dst1, block_offset, block, stride, nnzc, cntr, coeff, dst2, picreg
mov r5, 16 mov r5, 16
add r2, 512 add r2, 512
...@@ -802,8 +809,9 @@ h264_add8x4_idct_sse2: ...@@ -802,8 +809,9 @@ h264_add8x4_idct_sse2:
%endif %endif
%endmacro %endmacro
; ff_h264_idct_add16_sse2(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16_8_sse2(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16_8, 5, 5 + ARCH_X86_64, 8 cglobal h264_idct_add16_8, 5, 5 + ARCH_X86_64, 8
%if ARCH_X86_64 %if ARCH_X86_64
mov r5, r0 mov r5, r0
...@@ -849,8 +857,9 @@ cglobal h264_idct_add16_8, 5, 5 + ARCH_X86_64, 8 ...@@ -849,8 +857,9 @@ cglobal h264_idct_add16_8, 5, 5 + ARCH_X86_64, 8
%endif %endif
%endmacro %endmacro
; ff_h264_idct_add16intra_sse2(uint8_t *dst, const int *block_offset, ; ff_h264_idct_add16intra_8_sse2(uint8_t *dst, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add16intra_8, 5, 7 + ARCH_X86_64, 8 cglobal h264_idct_add16intra_8, 5, 7 + ARCH_X86_64, 8
%if ARCH_X86_64 %if ARCH_X86_64
mov r7, r0 mov r7, r0
...@@ -900,8 +909,9 @@ cglobal h264_idct_add16intra_8, 5, 7 + ARCH_X86_64, 8 ...@@ -900,8 +909,9 @@ cglobal h264_idct_add16intra_8, 5, 7 + ARCH_X86_64, 8
%endif %endif
%endmacro %endmacro
; ff_h264_idct_add8_sse2(uint8_t **dest, const int *block_offset, ; ff_h264_idct_add8_8_sse2(uint8_t **dest, const int *block_offset,
; int16_t *block, int stride, const uint8_t nnzc[6*8]) ; int16_t *block, int stride,
; const uint8_t nnzc[6 * 8])
cglobal h264_idct_add8_8, 5, 7 + ARCH_X86_64, 8 cglobal h264_idct_add8_8, 5, 7 + ARCH_X86_64, 8
add r2, 512 add r2, 512
%if ARCH_X86_64 %if ARCH_X86_64
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment