1/* Copyright (c) 2014, Cisco Systems, INC
2   Written by XiangMingZhu WeiZhou MinPeng YanWang
3
4   Redistribution and use in source and binary forms, with or without
5   modification, are permitted provided that the following conditions
6   are met:
7
8   - Redistributions of source code must retain the above copyright
9   notice, this list of conditions and the following disclaimer.
10
11   - Redistributions in binary form must reproduce the above copyright
12   notice, this list of conditions and the following disclaimer in the
13   documentation and/or other materials provided with the distribution.
14
15   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
16   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
17   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
18   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
19   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
20   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
21   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
22   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
23   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
24   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26*/
27
28#ifdef HAVE_CONFIG_H
29#include "config.h"
30#endif
31
32#include <xmmintrin.h>
33#include <emmintrin.h>
34#include <smmintrin.h>
35#include "celt_lpc.h"
36#include "stack_alloc.h"
37#include "mathops.h"
38#include "pitch.h"
39#include "x86cpu.h"
40
41#if defined(FIXED_POINT)
42
43void celt_fir_sse4_1(const opus_val16 *_x,
44         const opus_val16 *num,
45         opus_val16 *_y,
46         int N,
47         int ord,
48         opus_val16 *mem,
49         int arch)
50{
51    int i,j;
52    VARDECL(opus_val16, rnum);
53    VARDECL(opus_val16, x);
54
55    __m128i vecNoA;
56    opus_int32 noA ;
57    SAVE_STACK;
58
59   ALLOC(rnum, ord, opus_val16);
60   ALLOC(x, N+ord, opus_val16);
61   for(i=0;i<ord;i++)
62      rnum[i] = num[ord-i-1];
63   for(i=0;i<ord;i++)
64      x[i] = mem[ord-i-1];
65
66   for (i=0;i<N-7;i+=8)
67   {
68       x[i+ord  ]=_x[i  ];
69       x[i+ord+1]=_x[i+1];
70       x[i+ord+2]=_x[i+2];
71       x[i+ord+3]=_x[i+3];
72       x[i+ord+4]=_x[i+4];
73       x[i+ord+5]=_x[i+5];
74       x[i+ord+6]=_x[i+6];
75       x[i+ord+7]=_x[i+7];
76   }
77
78   for (;i<N-3;i+=4)
79   {
80       x[i+ord  ]=_x[i  ];
81       x[i+ord+1]=_x[i+1];
82       x[i+ord+2]=_x[i+2];
83       x[i+ord+3]=_x[i+3];
84   }
85
86   for (;i<N;i++)
87         x[i+ord]=_x[i];
88
89   for(i=0;i<ord;i++)
90      mem[i] = _x[N-i-1];
91#ifdef SMALL_FOOTPRINT
92   for (i=0;i<N;i++)
93   {
94      opus_val32 sum = SHL32(EXTEND32(_x[i]), SIG_SHIFT);
95      for (j=0;j<ord;j++)
96      {
97         sum = MAC16_16(sum,rnum[j],x[i+j]);
98      }
99      _y[i] = SATURATE16(PSHR32(sum, SIG_SHIFT));
100   }
101#else
102   noA = EXTEND32(1) << SIG_SHIFT >> 1;
103   vecNoA = _mm_set_epi32(noA, noA, noA, noA);
104
105   for (i=0;i<N-3;i+=4)
106   {
107      opus_val32 sums[4] = {0};
108      __m128i vecSum, vecX;
109
110      xcorr_kernel(rnum, x+i, sums, ord, arch);
111
112      vecSum = _mm_loadu_si128((__m128i *)sums);
113      vecSum = _mm_add_epi32(vecSum, vecNoA);
114      vecSum = _mm_srai_epi32(vecSum, SIG_SHIFT);
115      vecX = OP_CVTEPI16_EPI32_M64(_x + i);
116      vecSum = _mm_add_epi32(vecSum, vecX);
117      vecSum = _mm_packs_epi32(vecSum, vecSum);
118      _mm_storel_epi64((__m128i *)(_y + i), vecSum);
119   }
120   for (;i<N;i++)
121   {
122      opus_val32 sum = 0;
123      for (j=0;j<ord;j++)
124         sum = MAC16_16(sum, rnum[j], x[i + j]);
125      _y[i] = SATURATE16(ADD32(EXTEND32(_x[i]), PSHR32(sum, SIG_SHIFT)));
126   }
127
128#endif
129   RESTORE_STACK;
130}
131
132#endif
133