yading@10: ;******************************************************************************
yading@10: ;* MMX optimized DSP utils
yading@10: ;* Copyright (c) 2008 Loren Merritt
yading@10: ;* Copyright (c) 2003-2013 Michael Niedermayer
yading@10: ;* Copyright (c) 2013 Daniel Kang
yading@10: ;*
yading@10: ;* This file is part of FFmpeg.
yading@10: ;*
yading@10: ;* FFmpeg is free software; you can redistribute it and/or
yading@10: ;* modify it under the terms of the GNU Lesser General Public
yading@10: ;* License as published by the Free Software Foundation; either
yading@10: ;* version 2.1 of the License, or (at your option) any later version.
yading@10: ;*
yading@10: ;* FFmpeg is distributed in the hope that it will be useful,
yading@10: ;* but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10: ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10: ;* Lesser General Public License for more details.
yading@10: ;*
yading@10: ;* You should have received a copy of the GNU Lesser General Public
yading@10: ;* License along with FFmpeg; if not, write to the Free Software
yading@10: ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10: ;******************************************************************************
yading@10: 
yading@10: %include "libavutil/x86/x86util.asm"
yading@10: 
yading@10: SECTION .text
yading@10: 
yading@10: INIT_MMX mmxext
yading@10: ; void pixels(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
yading@10: %macro PIXELS48 2
yading@10: %if %2 == 4
yading@10: %define OP movh
yading@10: %else
yading@10: %define OP mova
yading@10: %endif
yading@10: cglobal %1_pixels%2, 4,5
yading@10:     movsxdifnidn r2, r2d
yading@10:     lea          r4, [r2*3]
yading@10: .loop:
yading@10:     OP           m0, [r1]
yading@10:     OP           m1, [r1+r2]
yading@10:     OP           m2, [r1+r2*2]
yading@10:     OP           m3, [r1+r4]
yading@10:     lea          r1, [r1+r2*4]
yading@10: %ifidn %1, avg
yading@10:     pavgb        m0, [r0]
yading@10:     pavgb        m1, [r0+r2]
yading@10:     pavgb        m2, [r0+r2*2]
yading@10:     pavgb        m3, [r0+r4]
yading@10: %endif
yading@10:     OP         [r0], m0
yading@10:     OP      [r0+r2], m1
yading@10:     OP    [r0+r2*2], m2
yading@10:     OP      [r0+r4], m3
yading@10:     sub         r3d, 4
yading@10:     lea          r0, [r0+r2*4]
yading@10:     jne       .loop
yading@10:     RET
yading@10: %endmacro
yading@10: 
yading@10: PIXELS48 put, 4
yading@10: PIXELS48 avg, 4
yading@10: PIXELS48 put, 8
yading@10: PIXELS48 avg, 8
yading@10: 
yading@10: 
yading@10: INIT_XMM sse2
yading@10: ; void put_pixels16_sse2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
yading@10: cglobal put_pixels16, 4,5,4
yading@10:     lea          r4, [r2*3]
yading@10: .loop:
yading@10:     movu         m0, [r1]
yading@10:     movu         m1, [r1+r2]
yading@10:     movu         m2, [r1+r2*2]
yading@10:     movu         m3, [r1+r4]
yading@10:     lea          r1, [r1+r2*4]
yading@10:     mova       [r0], m0
yading@10:     mova    [r0+r2], m1
yading@10:     mova  [r0+r2*2], m2
yading@10:     mova    [r0+r4], m3
yading@10:     sub         r3d, 4
yading@10:     lea          r0, [r0+r2*4]
yading@10:     jnz       .loop
yading@10:     REP_RET
yading@10: 
yading@10: ; void avg_pixels16_sse2(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
yading@10: cglobal avg_pixels16, 4,5,4
yading@10:     lea          r4, [r2*3]
yading@10: .loop:
yading@10:     movu         m0, [r1]
yading@10:     movu         m1, [r1+r2]
yading@10:     movu         m2, [r1+r2*2]
yading@10:     movu         m3, [r1+r4]
yading@10:     lea          r1, [r1+r2*4]
yading@10:     pavgb        m0, [r0]
yading@10:     pavgb        m1, [r0+r2]
yading@10:     pavgb        m2, [r0+r2*2]
yading@10:     pavgb        m3, [r0+r4]
yading@10:     mova       [r0], m0
yading@10:     mova    [r0+r2], m1
yading@10:     mova  [r0+r2*2], m2
yading@10:     mova    [r0+r4], m3
yading@10:     sub         r3d, 4
yading@10:     lea          r0, [r0+r2*4]
yading@10:     jnz       .loop
yading@10:     REP_RET