1 /*!
2 **********************************************************************************
3 * Copyright (c) 2021 Loongson Technology Corporation Limited
4 * Contributed by Lu Wang <wanglu@loongson.cn>
5 *
6 * \copy
7 * Copyright (c) 2013, Cisco Systems
8 * All rights reserved.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 *
14 * * Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 *
17 * * Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in
19 * the documentation and/or other materials provided with the
20 * distribution.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
23 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
24 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
25 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
26 * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
27 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
28 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
29 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
30 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
32 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
33 * POSSIBILITY OF SUCH DAMAGE.
34 *
35 * \file mc_chroma_lsx.c
36 *
37 * \brief Loongson optimization
38 *
39 * \date 12/23/2021 Created
40 *
41 **********************************************************************************
42 */
43
44 #include "stdint.h"
45 #include "loongson_intrinsics.h"
46
McCopyWidthEq4_lsx(const uint8_t * pSrc,int iSrcStride,uint8_t * pDst,int iDstStride,int iHeight)47 void McCopyWidthEq4_lsx(const uint8_t *pSrc, int iSrcStride,
48 uint8_t *pDst, int iDstStride, int iHeight) {
49 __m128i src0, src1;
50 int iSrcStride_x2 = iSrcStride << 1;
51 int iDstStride_x2 = iDstStride << 1;
52 iHeight = iHeight >> 1;
53 for (int i = 0; i < iHeight; i++) {
54 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0, src1);
55 __lsx_vstelm_w(src0, pDst, 0, 0);
56 __lsx_vstelm_w(src1, pDst + iDstStride, 0, 0);
57 pSrc += iSrcStride_x2;
58 pDst += iDstStride_x2;
59 }
60 }
61
McCopyWidthEq8_lsx(const uint8_t * pSrc,int iSrcStride,uint8_t * pDst,int iDstStride,int iHeight)62 void McCopyWidthEq8_lsx(const uint8_t *pSrc, int iSrcStride,
63 uint8_t *pDst, int iDstStride, int iHeight) {
64 __m128i src0, src1;
65 int iSrcStride_x2 = iSrcStride << 1;
66 int iDstStride_x2 = iDstStride << 1;
67 iHeight = iHeight >> 1;
68 for (int i = 0; i < iHeight; i++) {
69 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0, src1);
70 __lsx_vstelm_d(src0, pDst, 0, 0);
71 __lsx_vstelm_d(src1, pDst + iDstStride, 0, 0);
72 pSrc += iSrcStride_x2;
73 pDst += iDstStride_x2;
74 }
75 }
76
McCopyWidthEq16_lsx(const uint8_t * pSrc,int iSrcStride,uint8_t * pDst,int iDstStride,int iHeight)77 void McCopyWidthEq16_lsx(const uint8_t *pSrc, int iSrcStride,
78 uint8_t *pDst, int iDstStride, int iHeight) {
79 __m128i src0, src1;
80 int iSrcStride_x2 = iSrcStride << 1;
81 int iDstStride_x2 = iDstStride << 1;
82 iHeight = iHeight >> 1;
83 for (int i = 0; i < iHeight; i++) {
84 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0, src1);
85 __lsx_vst(src0, pDst, 0);
86 __lsx_vstx(src1, pDst, iDstStride);
87 pSrc += iSrcStride_x2;
88 pDst += iDstStride_x2;
89 }
90 }
91
McChromaWidthEq4_lsx(const uint8_t * pSrc,int32_t iSrcStride,uint8_t * pDst,int32_t iDstStride,const uint8_t * pABCD,int32_t iHeight)92 void McChromaWidthEq4_lsx(const uint8_t *pSrc, int32_t iSrcStride, uint8_t *pDst,
93 int32_t iDstStride, const uint8_t *pABCD, int32_t iHeight) {
94 int32_t i;
95 int8_t iA, iB, iC, iD;
96 uint16_t iFillData;
97 int32_t iSrcStride_x2 = iSrcStride << 1;
98 int32_t iDstStride_x2 = iDstStride << 1;
99 __m128i paramAB, paramCD, paramConst;
100 __m128i src0, src1, src2, src0Even, src1Even, src2Even;
101 __m128i out0, outNext0, out1, outNext1;
102 __m128i shift;
103
104 iA = pABCD[0];
105 iB = pABCD[1];
106 iC = pABCD[2];
107 iD = pABCD[3];
108
109 iFillData = (iB << 8) | iA;
110 paramAB = __lsx_vreplgr2vr_h(iFillData);
111 iFillData = (iD << 8) | iC;
112 paramCD = __lsx_vreplgr2vr_h(iFillData);
113 paramConst = __lsx_vreplgr2vr_h(32);
114 shift = __lsx_vreplgr2vr_h(6);
115 iHeight = iHeight >> 1;
116
117 for (i = 0; i < iHeight; i++) {
118 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0, src1);
119 src2 = __lsx_vld(pSrc + iSrcStride_x2, 0);
120 pSrc += 1;
121 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0Even, src1Even);
122 src2Even = __lsx_vld(pSrc + iSrcStride_x2, 0);
123 pSrc -= 1;
124
125 DUP2_ARG2(__lsx_vilvl_w, src0Even, src0, src1Even, src1, src0, src1);
126 src2 = __lsx_vilvl_w(src2Even, src2);
127 DUP4_ARG2(__lsx_vdp2_h_bu, src0, paramAB, src1, paramCD, src1, paramAB, src2,
128 paramCD, out0, outNext0, out1, outNext1);
129 DUP4_ARG2(__lsx_vadd_h, out0, outNext0, out0, paramConst, out1, outNext1, out1,
130 paramConst, out0, out0, out1, out1);
131 DUP2_ARG2(__lsx_vsrl_h, out0, shift, out1, shift, out0, out1);
132 DUP2_ARG2(__lsx_vsrl_d, out0, paramConst, out1, paramConst, outNext0, outNext1);
133 DUP2_ARG2(__lsx_vpackev_b, outNext0, out0, outNext1, out1, out0, out1);
134 __lsx_vstelm_w(out0, pDst, 0, 0);
135 __lsx_vstelm_w(out1, pDst + iDstStride, 0, 0);
136
137 pDst += iDstStride_x2;
138 pSrc += iSrcStride_x2;
139 }
140 }
141
McChromaWidthEq8_lsx(const uint8_t * pSrc,int32_t iSrcStride,uint8_t * pDst,int32_t iDstStride,const uint8_t * pABCD,int32_t iHeight)142 void McChromaWidthEq8_lsx(const uint8_t *pSrc, int32_t iSrcStride, uint8_t *pDst,
143 int32_t iDstStride, const uint8_t *pABCD, int32_t iHeight) {
144 int32_t i;
145 int8_t iA, iB, iC, iD;
146 uint16_t iFillData;
147 int32_t iSrcStride_x2 = iSrcStride << 1;
148 int32_t iDstStride_x2 = iDstStride << 1;
149 __m128i paramAB, paramCD, paramConst;
150 __m128i src0, src1, src2, src0Even, src1Even, src2Even;
151 __m128i out0, outNext0, out1, outNext1;
152 __m128i shift;
153
154 iA = pABCD[0];
155 iB = pABCD[1];
156 iC = pABCD[2];
157 iD = pABCD[3];
158
159 iFillData = (iB << 8) | iA;
160 paramAB = __lsx_vreplgr2vr_h(iFillData);
161 iFillData = (iD << 8) | iC;
162 paramCD = __lsx_vreplgr2vr_h(iFillData);
163 paramConst = __lsx_vreplgr2vr_h(32);
164 shift = __lsx_vreplgr2vr_h(6);
165 iHeight = iHeight >> 1;
166
167 for (i = 0; i < iHeight; i++) {
168 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0, src1);
169 src2 = __lsx_vld(pSrc + iSrcStride_x2, 0);
170 pSrc += 1;
171 DUP2_ARG2(__lsx_vld, pSrc, 0, pSrc + iSrcStride, 0, src0Even, src1Even);
172 src2Even = __lsx_vld(pSrc + iSrcStride_x2, 0);
173 pSrc -= 1;
174
175 DUP2_ARG2(__lsx_vilvl_d, src0Even, src0, src1Even, src1, src0, src1);
176 src2 = __lsx_vilvl_d(src2Even, src2);
177 DUP4_ARG2(__lsx_vdp2_h_bu, src0, paramAB, src1, paramCD, src1, paramAB, src2,
178 paramCD, out0, outNext0, out1, outNext1);
179 DUP4_ARG2(__lsx_vadd_h, out0, outNext0, out0, paramConst, out1, outNext1, out1,
180 paramConst, out0, out0, out1, out1);
181 DUP2_ARG2(__lsx_vsrl_h, out0, shift, out1, shift, out0, out1);
182
183 DUP2_ARG2(__lsx_vilvh_d, out0, out0, out1, out1, outNext0, outNext1);
184 DUP2_ARG2(__lsx_vpackev_b, outNext0, out0, outNext1, out1, out0, out1);
185
186 __lsx_vstelm_d(out0, pDst, 0, 0);
187 __lsx_vstelm_d(out1, pDst + iDstStride, 0, 0);
188
189 pDst += iDstStride_x2;
190 pSrc += iSrcStride_x2;
191 }
192 }
193