1 ;******************************************************************************
2 ;* optimized audio functions
3 ;* Copyright (c) 2008 Loren Merritt
5 ;* This file is part of Libav.
7 ;* Libav is free software; you can redistribute it and/or
8 ;* modify it under the terms of the GNU Lesser General Public
9 ;* License as published by the Free Software Foundation; either
10 ;* version 2.1 of the License, or (at your option) any later version.
12 ;* Libav is distributed in the hope that it will be useful,
13 ;* but WITHOUT ANY WARRANTY; without even the implied warranty of
14 ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 ;* Lesser General Public License for more details.
17 ;* You should have received a copy of the GNU Lesser General Public
18 ;* License along with Libav; if not, write to the Free Software
19 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 ;******************************************************************************
22 %include "libavutil/x86/x86util.asm"
26 %macro SCALARPRODUCT 0
27 ; int ff_scalarproduct_int16(int16_t *v1, int16_t *v2, int order)
28 cglobal scalarproduct_int16, 3,3,3, v1, v2, order
35 movu m0, [v1q + orderq]
36 movu m1, [v1q + orderq + mmsize]
37 pmaddwd m0, [v2q + orderq]
38 pmaddwd m1, [v2q + orderq + mmsize]
61 ;-----------------------------------------------------------------------------
62 ; void ff_vector_clip_int32(int32_t *dst, const int32_t *src, int32_t min,
63 ; int32_t max, unsigned int len)
64 ;-----------------------------------------------------------------------------
66 ; %1 = number of xmm registers used
67 ; %2 = number of inline load/process/store loops per asm loop
68 ; %3 = process 4*mmsize (%3=0) or 8*mmsize (%3=1) bytes per loop
69 ; %4 = CLIPD function takes min/max as float instead of int (CLIPD_SSE2)
71 %macro VECTOR_CLIP_INT32 4-5
72 cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
85 mova m0, [srcq + mmsize * (0 + %%i)]
86 mova m1, [srcq + mmsize * (1 + %%i)]
87 mova m2, [srcq + mmsize * (2 + %%i)]
88 mova m3, [srcq + mmsize * (3 + %%i)]
90 mova m7, [srcq + mmsize * (4 + %%i)]
91 mova m8, [srcq + mmsize * (5 + %%i)]
92 mova m9, [srcq + mmsize * (6 + %%i)]
93 mova m10, [srcq + mmsize * (7 + %%i)]
103 CLIPD m10, m4, m5, m6
105 mova [dstq + mmsize * (0 + %%i)], m0
106 mova [dstq + mmsize * (1 + %%i)], m1
107 mova [dstq + mmsize * (2 + %%i)], m2
108 mova [dstq + mmsize * (3 + %%i)], m3
110 mova [dstq + mmsize * (4 + %%i)], m7
111 mova [dstq + mmsize * (5 + %%i)], m8
112 mova [dstq + mmsize * (6 + %%i)], m9
113 mova [dstq + mmsize * (7 + %%i)], m10
115 %assign %%i (%%i + 4 * (1 + %3))
117 add srcq, mmsize*4*(%2+%3)
118 add dstq, mmsize*4*(%2+%3)
119 sub lend, mmsize*(%2+%3)
125 %define CLIPD CLIPD_MMX
126 VECTOR_CLIP_INT32 0, 1, 0, 0
128 VECTOR_CLIP_INT32 6, 1, 0, 0, _int
129 %define CLIPD CLIPD_SSE2
130 VECTOR_CLIP_INT32 6, 2, 0, 1
132 %define CLIPD CLIPD_SSE41
134 VECTOR_CLIP_INT32 11, 1, 1, 0
136 VECTOR_CLIP_INT32 6, 1, 0, 0
139 ; void ff_vector_clipf_sse(float *dst, const float *src,
140 ; int len, float min, float max)
142 cglobal vector_clipf, 3, 3, 6, dst, src, len, min, max
144 VBROADCASTSS m0, minm
145 VBROADCASTSS m1, maxm
148 VBROADCASTSS m1, maxm
154 movsxdifnidn lenq, lend
157 mova m2, [srcq + 4 * lenq - 4 * mmsize]
158 mova m3, [srcq + 4 * lenq - 3 * mmsize]
159 mova m4, [srcq + 4 * lenq - 2 * mmsize]
160 mova m5, [srcq + 4 * lenq - 1 * mmsize]
172 mova [dstq + 4 * lenq - 4 * mmsize], m2
173 mova [dstq + 4 * lenq - 3 * mmsize], m3
174 mova [dstq + 4 * lenq - 2 * mmsize], m4
175 mova [dstq + 4 * lenq - 1 * mmsize], m5