diff options
| author | Christophe Gisquet <christophe.gisquet@gmail.com> | 2012-05-11 11:25:30 +0200 |
|---|---|---|
| committer | Janne Grunau <janne-libav@jannau.net> | 2014-02-07 22:52:40 +0100 |
| commit | 5b59a9fc6152169599561f04b4f66370edda5c9c (patch) | |
| tree | ae7358874208c0c6880baf9c7132dd5fb1bc6262 /libavcodec/x86/dcadsp.asm | |
| parent | 2bd44cb705340c4f7bd7e459a1efed5074bf45fc (diff) | |
x86: dcadsp: implement int8x8_fmul_int32
For the callable function (as opposed to the inline one):
C SSE SSE2 SSE4
Win32: 47 42 29 26
Win64: 30 33 25 23
The SSE version is neither compiled nor set for ARCH_X86_64, as the
inlinable function takes over.
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
Diffstat (limited to 'libavcodec/x86/dcadsp.asm')
| -rw-r--r-- | libavcodec/x86/dcadsp.asm | 90 |
1 files changed, 90 insertions, 0 deletions
diff --git a/libavcodec/x86/dcadsp.asm b/libavcodec/x86/dcadsp.asm new file mode 100644 index 0000000000..214f5146aa --- /dev/null +++ b/libavcodec/x86/dcadsp.asm @@ -0,0 +1,90 @@ +;****************************************************************************** +;* SSE-optimized functions for the DCA decoder +;* Copyright (C) 2012-2014 Christophe Gisquet <christophe.gisquet@gmail.com> +;* +;* This file is part of Libav. +;* +;* Libav is free software; you can redistribute it and/or +;* modify it under the terms of the GNU Lesser General Public +;* License as published by the Free Software Foundation; either +;* version 2.1 of the License, or (at your option) any later version. +;* +;* Libav 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 +;* Lesser General Public License for more details. +;* +;* You should have received a copy of the GNU Lesser General Public +;* License along with Libav; if not, write to the Free Software +;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +;****************************************************************************** + +%include "libavutil/x86/x86util.asm" + +SECTION_RODATA +pf_inv16: times 4 dd 0x3D800000 ; 1/16 + +SECTION_TEXT + +; void int8x8_fmul_int32_sse2(float *dst, const int8_t *src, int scale) +%macro INT8X8_FMUL_INT32 0 +cglobal int8x8_fmul_int32, 3,3,5, dst, src, scale + cvtsi2ss m0, scalem + mulss m0, [pf_inv16] + shufps m0, m0, 0 +%if cpuflag(sse2) +%if cpuflag(sse4) + pmovsxbd m1, [srcq+0] + pmovsxbd m2, [srcq+4] +%else + movq m1, [srcq] + punpcklbw m1, m1 + mova m2, m1 + punpcklwd m1, m1 + punpckhwd m2, m2 + psrad m1, 24 + psrad m2, 24 +%endif + cvtdq2ps m1, m1 + cvtdq2ps m2, m2 +%else + movd mm0, [srcq+0] + movd mm1, [srcq+4] + punpcklbw mm0, mm0 + punpcklbw mm1, mm1 + movq mm2, mm0 + movq mm3, mm1 + punpcklwd mm0, mm0 + punpcklwd mm1, mm1 + punpckhwd mm2, mm2 + punpckhwd mm3, mm3 + psrad mm0, 24 + psrad mm1, 24 + psrad mm2, 24 + psrad mm3, 24 + cvtpi2ps m1, mm0 + cvtpi2ps m2, mm1 + cvtpi2ps m3, mm2 + cvtpi2ps m4, mm3 + shufps m0, m0, 0 + emms + shufps m1, m3, q1010 + shufps m2, m4, q1010 +%endif + mulps m1, m0 + mulps m2, m0 + mova [dstq+ 0], m1 + mova [dstq+16], m2 + REP_RET +%endmacro + +%if ARCH_X86_32 +INIT_XMM sse +INT8X8_FMUL_INT32 +%endif + +INIT_XMM sse2 +INT8X8_FMUL_INT32 + +INIT_XMM sse4 +INT8X8_FMUL_INT32 |