1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
|
/*
* The ManaPlus Client
* Copyright (C) 2007-2009 The Mana World Development Team
* Copyright (C) 2009-2010 The Mana Developers
* Copyright (C) 2011-2020 The ManaPlus Developers
*
* This file is part of The ManaPlus Client.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* any later version.
*
* This program 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "resources/dye/dyepalette.h"
PRAGMA48(GCC diagnostic push)
PRAGMA48(GCC diagnostic ignored "-Wshadow")
#ifndef SDL_BIG_ENDIAN
#include <SDL_endian.h>
#endif // SDL_BYTEORDER
PRAGMA48(GCC diagnostic pop)
#ifdef SIMD_SUPPORTED
// avx2
#include <immintrin.h>
#endif // SIMD_SUPPORTED
#include "debug.h"
void DyePalette::replaceAColorDefault(uint32_t *restrict pixels,
const int bufSize) const restrict2
{
STD_VECTOR<DyeColor>::const_iterator it_end = mColors.end();
const size_t sz = mColors.size();
if ((sz == 0U) || (pixels == nullptr))
return;
if ((sz % 2) != 0U)
-- it_end;
for (const uint32_t *const p_end = pixels + CAST_SIZE(bufSize);
pixels != p_end;
++pixels)
{
uint8_t *const p = reinterpret_cast<uint8_t *>(pixels);
const unsigned int data = *pixels;
STD_VECTOR<DyeColor>::const_iterator it = mColors.begin();
while (it != it_end)
{
const DyeColor &col = *it;
++ it;
const DyeColor &col2 = *it;
#if SDL_BYTEORDER == SDL_BIG_ENDIAN
const unsigned int coldata = (col.value[3] << 24U)
| (col.value[2] << 16U)
| (col.value[1] << 8U)
| (col.value[0]);
#else // SDL_BYTEORDER == SDL_BIG_ENDIAN
const unsigned int coldata = (col.value[3])
| (col.value[2] << 8U)
| (col.value[1] << 16U) |
(col.value[0] << 24U);
#endif // SDL_BYTEORDER == SDL_BIG_ENDIAN
if (data == coldata)
{
p[3] = col2.value[0];
p[2] = col2.value[1];
p[1] = col2.value[2];
p[0] = col2.value[3];
break;
}
++ it;
}
}
}
#ifdef SIMD_SUPPORTED
/*
static void print256(const char *const text, const __m256i &val);
static void print256(const char *const text, const __m256i &val)
{
printf("%s 0x%016llx%016llx%016llx%016llx\n", text, val[0], val[1], val[2], val[3]);
}
*/
__attribute__ ((target ("sse2")))
void DyePalette::replaceAColorSse2(uint32_t *restrict pixels,
const int bufSize) const restrict2
{
STD_VECTOR<DyeColor>::const_iterator it_end = mColors.end();
const size_t sz = mColors.size();
if ((sz == 0U) || (pixels == nullptr))
return;
if ((sz % 2) != 0U)
-- it_end;
const int mod = bufSize % 4;
const int bufEnd = bufSize - mod;
for (int ptr = 0; ptr < bufEnd; ptr += 4)
{
// __m128i base = _mm_load_si128(reinterpret_cast<__m128i*>(pixels));
__m128i base = _mm_loadu_si128(reinterpret_cast<__m128i*>(
&pixels[ptr]));
STD_VECTOR<DyeColor>::const_iterator it = mColors.begin();
while (it != it_end)
{
const DyeColor &col = *it;
++ it;
const DyeColor &col2 = *it;
__m128i newMask = _mm_set1_epi32(col2.valueA);
__m128i cmpMask = _mm_set1_epi32(col.valueA);
__m128i cmpRes = _mm_cmpeq_epi32(base, cmpMask);
__m128i srcAnd = _mm_andnot_si128(cmpRes, base);
__m128i dstAnd = _mm_and_si128(cmpRes, newMask);
base = _mm_or_si128(srcAnd, dstAnd);
++ it;
}
// _mm_store_si128(reinterpret_cast<__m128i*>(pixels), base);
_mm_storeu_si128(reinterpret_cast<__m128i*>(&pixels[ptr]), base);
}
// complete end without simd
for (int ptr = bufSize - mod; ptr < bufSize; ptr ++)
{
uint8_t *const p = reinterpret_cast<uint8_t *>(&pixels[ptr]);
const unsigned int data = pixels[ptr];
STD_VECTOR<DyeColor>::const_iterator it = mColors.begin();
while (it != it_end)
{
const DyeColor &col = *it;
++ it;
const DyeColor &col2 = *it;
const unsigned int coldata = (col.value[3]) |
(col.value[2] << 8U) |
(col.value[1] << 16U) |
(col.value[0] << 24U);
if (data == coldata)
{
p[3] = col2.value[0];
p[2] = col2.value[1];
p[1] = col2.value[2];
p[0] = col2.value[3];
break;
}
++ it;
}
}
}
__attribute__ ((target ("avx2")))
void DyePalette::replaceAColorAvx2(uint32_t *restrict pixels,
const int bufSize) const restrict2
{
STD_VECTOR<DyeColor>::const_iterator it_end = mColors.end();
const size_t sz = mColors.size();
if ((sz == 0U) || (pixels == nullptr))
return;
if ((sz % 2) != 0U)
-- it_end;
const int mod = bufSize % 8;
const int bufEnd = bufSize - mod;
for (int ptr = 0; ptr < bufEnd; ptr += 8)
{
// __m256i base = _mm256_load_si256(reinterpret_cast<__m256i*>(pixels));
__m256i base = _mm256_loadu_si256(reinterpret_cast<__m256i*>(
&pixels[ptr]));
STD_VECTOR<DyeColor>::const_iterator it = mColors.begin();
while (it != it_end)
{
const DyeColor &col = *it;
++ it;
const DyeColor &col2 = *it;
__m256i newMask = _mm256_set1_epi32(col2.valueA);
__m256i cmpMask = _mm256_set1_epi32(col.valueA);
__m256i cmpRes = _mm256_cmpeq_epi32(base, cmpMask);
__m256i srcAnd = _mm256_andnot_si256(cmpRes, base);
__m256i dstAnd = _mm256_and_si256(cmpRes, newMask);
base = _mm256_or_si256(srcAnd, dstAnd);
++ it;
}
// _mm256_store_si256(reinterpret_cast<__m256i*>(pixels), base);
_mm256_storeu_si256(reinterpret_cast<__m256i*>(&pixels[ptr]), base);
}
// complete end without simd
for (int ptr = bufSize - mod; ptr < bufSize; ptr ++)
{
uint8_t *const p = reinterpret_cast<uint8_t *>(&pixels[ptr]);
const unsigned int data = pixels[ptr];
STD_VECTOR<DyeColor>::const_iterator it = mColors.begin();
while (it != it_end)
{
const DyeColor &col = *it;
++ it;
const DyeColor &col2 = *it;
const unsigned int coldata = (col.value[3]) |
(col.value[2] << 8U) |
(col.value[1] << 16U) |
(col.value[0] << 24U);
if (data == coldata)
{
p[3] = col2.value[0];
p[2] = col2.value[1];
p[1] = col2.value[2];
p[0] = col2.value[3];
break;
}
++ it;
}
}
}
#endif // SIMD_SUPPORTED
|
