1 | /**************************************************************************** |
2 | ** |
3 | ** Copyright (C) 2016 The Qt Company Ltd. |
4 | ** Contact: https://www.qt.io/licensing/ |
5 | ** |
6 | ** This file is part of the QtGui module of the Qt Toolkit. |
7 | ** |
8 | ** $QT_BEGIN_LICENSE:LGPL$ |
9 | ** Commercial License Usage |
10 | ** Licensees holding valid commercial Qt licenses may use this file in |
11 | ** accordance with the commercial license agreement provided with the |
12 | ** Software or, alternatively, in accordance with the terms contained in |
13 | ** a written agreement between you and The Qt Company. For licensing terms |
14 | ** and conditions see https://www.qt.io/terms-conditions. For further |
15 | ** information use the contact form at https://www.qt.io/contact-us. |
16 | ** |
17 | ** GNU Lesser General Public License Usage |
18 | ** Alternatively, this file may be used under the terms of the GNU Lesser |
19 | ** General Public License version 3 as published by the Free Software |
20 | ** Foundation and appearing in the file LICENSE.LGPL3 included in the |
21 | ** packaging of this file. Please review the following information to |
22 | ** ensure the GNU Lesser General Public License version 3 requirements |
23 | ** will be met: https://www.gnu.org/licenses/lgpl-3.0.html. |
24 | ** |
25 | ** GNU General Public License Usage |
26 | ** Alternatively, this file may be used under the terms of the GNU |
27 | ** General Public License version 2.0 or (at your option) the GNU General |
28 | ** Public license version 3 or any later version approved by the KDE Free |
29 | ** Qt Foundation. The licenses are as published by the Free Software |
30 | ** Foundation and appearing in the file LICENSE.GPL2 and LICENSE.GPL3 |
31 | ** included in the packaging of this file. Please review the following |
32 | ** information to ensure the GNU General Public License requirements will |
33 | ** be met: https://www.gnu.org/licenses/gpl-2.0.html and |
34 | ** https://www.gnu.org/licenses/gpl-3.0.html. |
35 | ** |
36 | ** $QT_END_LICENSE$ |
37 | ** |
38 | ****************************************************************************/ |
39 | |
40 | #include <private/qdrawhelper_p.h> |
41 | #include <private/qdrawingprimitive_sse2_p.h> |
42 | #include <private/qpaintengine_raster_p.h> |
43 | |
44 | #if defined(QT_COMPILER_SUPPORTS_SSE4_1) |
45 | |
46 | QT_BEGIN_NAMESPACE |
47 | |
48 | #ifndef __AVX2__ |
49 | template<bool RGBA> |
50 | static void convertARGBToARGB32PM_sse4(uint *buffer, const uint *src, int count) |
51 | { |
52 | int i = 0; |
53 | const __m128i alphaMask = _mm_set1_epi32(i: 0xff000000); |
54 | const __m128i rgbaMask = _mm_setr_epi8(b0: 2, b1: 1, b2: 0, b3: 3, b4: 6, b5: 5, b6: 4, b7: 7, b8: 10, b9: 9, b10: 8, b11: 11, b12: 14, b13: 13, b14: 12, b15: 15); |
55 | const __m128i shuffleMask = _mm_setr_epi8(b0: 6, b1: 7, b2: 6, b3: 7, b4: 6, b5: 7, b6: 6, b7: 7, b8: 14, b9: 15, b10: 14, b11: 15, b12: 14, b13: 15, b14: 14, b15: 15); |
56 | const __m128i half = _mm_set1_epi16(w: 0x0080); |
57 | const __m128i zero = _mm_setzero_si128(); |
58 | |
59 | for (; i < count - 3; i += 4) { |
60 | __m128i srcVector = _mm_loadu_si128(p: (const __m128i *)&src[i]); |
61 | if (!_mm_testz_si128(M: srcVector, V: alphaMask)) { |
62 | if (!_mm_testc_si128(M: srcVector, V: alphaMask)) { |
63 | if (RGBA) |
64 | srcVector = _mm_shuffle_epi8(a: srcVector, b: rgbaMask); |
65 | __m128i src1 = _mm_unpacklo_epi8(a: srcVector, b: zero); |
66 | __m128i src2 = _mm_unpackhi_epi8(a: srcVector, b: zero); |
67 | __m128i alpha1 = _mm_shuffle_epi8(a: src1, b: shuffleMask); |
68 | __m128i alpha2 = _mm_shuffle_epi8(a: src2, b: shuffleMask); |
69 | src1 = _mm_mullo_epi16(a: src1, b: alpha1); |
70 | src2 = _mm_mullo_epi16(a: src2, b: alpha2); |
71 | src1 = _mm_add_epi16(a: src1, b: _mm_srli_epi16(a: src1, count: 8)); |
72 | src2 = _mm_add_epi16(a: src2, b: _mm_srli_epi16(a: src2, count: 8)); |
73 | src1 = _mm_add_epi16(a: src1, b: half); |
74 | src2 = _mm_add_epi16(a: src2, b: half); |
75 | src1 = _mm_srli_epi16(a: src1, count: 8); |
76 | src2 = _mm_srli_epi16(a: src2, count: 8); |
77 | src1 = _mm_blend_epi16(src1, alpha1, 0x88); |
78 | src2 = _mm_blend_epi16(src2, alpha2, 0x88); |
79 | srcVector = _mm_packus_epi16(a: src1, b: src2); |
80 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: srcVector); |
81 | } else { |
82 | if (RGBA) |
83 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: _mm_shuffle_epi8(a: srcVector, b: rgbaMask)); |
84 | else if (buffer != src) |
85 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: srcVector); |
86 | } |
87 | } else { |
88 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: zero); |
89 | } |
90 | } |
91 | |
92 | SIMD_EPILOGUE(i, count, 3) { |
93 | uint v = qPremultiply(x: src[i]); |
94 | buffer[i] = RGBA ? RGBA2ARGB(x: v) : v; |
95 | } |
96 | } |
97 | |
98 | template<bool RGBA> |
99 | static void convertARGBToRGBA64PM_sse4(QRgba64 *buffer, const uint *src, int count) |
100 | { |
101 | int i = 0; |
102 | const __m128i alphaMask = _mm_set1_epi32(i: 0xff000000); |
103 | const __m128i rgbaMask = _mm_setr_epi8(b0: 2, b1: 1, b2: 0, b3: 3, b4: 6, b5: 5, b6: 4, b7: 7, b8: 10, b9: 9, b10: 8, b11: 11, b12: 14, b13: 13, b14: 12, b15: 15); |
104 | const __m128i shuffleMask = _mm_setr_epi8(b0: 6, b1: 7, b2: 6, b3: 7, b4: 6, b5: 7, b6: 6, b7: 7, b8: 14, b9: 15, b10: 14, b11: 15, b12: 14, b13: 15, b14: 14, b15: 15); |
105 | const __m128i zero = _mm_setzero_si128(); |
106 | |
107 | for (; i < count - 3; i += 4) { |
108 | __m128i srcVector = _mm_loadu_si128(p: (const __m128i *)&src[i]); |
109 | if (!_mm_testz_si128(M: srcVector, V: alphaMask)) { |
110 | bool cf = _mm_testc_si128(M: srcVector, V: alphaMask); |
111 | |
112 | if (!RGBA) |
113 | srcVector = _mm_shuffle_epi8(a: srcVector, b: rgbaMask); |
114 | const __m128i src1 = _mm_unpacklo_epi8(a: srcVector, b: srcVector); |
115 | const __m128i src2 = _mm_unpackhi_epi8(a: srcVector, b: srcVector); |
116 | if (!cf) { |
117 | __m128i alpha1 = _mm_shuffle_epi8(a: src1, b: shuffleMask); |
118 | __m128i alpha2 = _mm_shuffle_epi8(a: src2, b: shuffleMask); |
119 | __m128i dst1 = _mm_mulhi_epu16(a: src1, b: alpha1); |
120 | __m128i dst2 = _mm_mulhi_epu16(a: src2, b: alpha2); |
121 | // Map 0->0xfffe to 0->0xffff |
122 | dst1 = _mm_add_epi16(a: dst1, b: _mm_srli_epi16(a: dst1, count: 15)); |
123 | dst2 = _mm_add_epi16(a: dst2, b: _mm_srli_epi16(a: dst2, count: 15)); |
124 | // correct alpha value: |
125 | dst1 = _mm_blend_epi16(dst1, src1, 0x88); |
126 | dst2 = _mm_blend_epi16(dst2, src2, 0x88); |
127 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: dst1); |
128 | _mm_storeu_si128(p: (__m128i *)&buffer[i + 2], b: dst2); |
129 | } else { |
130 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: src1); |
131 | _mm_storeu_si128(p: (__m128i *)&buffer[i + 2], b: src2); |
132 | } |
133 | } else { |
134 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: zero); |
135 | _mm_storeu_si128(p: (__m128i *)&buffer[i + 2], b: zero); |
136 | } |
137 | } |
138 | |
139 | SIMD_EPILOGUE(i, count, 3) { |
140 | const uint s = RGBA ? RGBA2ARGB(x: src[i]) : src[i]; |
141 | buffer[i] = QRgba64::fromArgb32(rgb: s).premultiplied(); |
142 | } |
143 | } |
144 | #endif // __AVX2__ |
145 | |
146 | static inline __m128 Q_DECL_VECTORCALL reciprocal_mul_ps(__m128 a, float mul) |
147 | { |
148 | __m128 ia = _mm_rcp_ps(a: a); // Approximate 1/a |
149 | // Improve precision of ia using Newton-Raphson |
150 | ia = _mm_sub_ps(a: _mm_add_ps(a: ia, b: ia), b: _mm_mul_ps(a: ia, b: _mm_mul_ps(a: ia, b: a))); |
151 | ia = _mm_mul_ps(a: ia, b: _mm_set1_ps(w: mul)); |
152 | return ia; |
153 | } |
154 | |
155 | template<bool RGBA, bool RGBx> |
156 | static inline void convertARGBFromARGB32PM_sse4(uint *buffer, const uint *src, int count) |
157 | { |
158 | int i = 0; |
159 | if ((_MM_GET_EXCEPTION_MASK() & _MM_MASK_INVALID) == 0) { |
160 | for (; i < count; ++i) { |
161 | uint v = qUnpremultiply(p: src[i]); |
162 | if (RGBx) |
163 | v = 0xff000000 | v; |
164 | if (RGBA) |
165 | v = ARGB2RGBA(x: v); |
166 | buffer[i] = v; |
167 | } |
168 | return; |
169 | } |
170 | const __m128i alphaMask = _mm_set1_epi32(i: 0xff000000); |
171 | const __m128i rgbaMask = _mm_setr_epi8(b0: 2, b1: 1, b2: 0, b3: 3, b4: 6, b5: 5, b6: 4, b7: 7, b8: 10, b9: 9, b10: 8, b11: 11, b12: 14, b13: 13, b14: 12, b15: 15); |
172 | const __m128i zero = _mm_setzero_si128(); |
173 | |
174 | for (; i < count - 3; i += 4) { |
175 | __m128i srcVector = _mm_loadu_si128(p: (const __m128i *)&src[i]); |
176 | if (!_mm_testz_si128(M: srcVector, V: alphaMask)) { |
177 | if (!_mm_testc_si128(M: srcVector, V: alphaMask)) { |
178 | __m128i srcVectorAlpha = _mm_srli_epi32(a: srcVector, count: 24); |
179 | if (RGBA) |
180 | srcVector = _mm_shuffle_epi8(a: srcVector, b: rgbaMask); |
181 | const __m128 a = _mm_cvtepi32_ps(a: srcVectorAlpha); |
182 | const __m128 ia = reciprocal_mul_ps(a, mul: 255.0f); |
183 | __m128i src1 = _mm_unpacklo_epi8(a: srcVector, b: zero); |
184 | __m128i src3 = _mm_unpackhi_epi8(a: srcVector, b: zero); |
185 | __m128i src2 = _mm_unpackhi_epi16(a: src1, b: zero); |
186 | __m128i src4 = _mm_unpackhi_epi16(a: src3, b: zero); |
187 | src1 = _mm_unpacklo_epi16(a: src1, b: zero); |
188 | src3 = _mm_unpacklo_epi16(a: src3, b: zero); |
189 | __m128 ia1 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(0, 0, 0, 0)); |
190 | __m128 ia2 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(1, 1, 1, 1)); |
191 | __m128 ia3 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(2, 2, 2, 2)); |
192 | __m128 ia4 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(3, 3, 3, 3)); |
193 | src1 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src1), b: ia1)); |
194 | src2 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src2), b: ia2)); |
195 | src3 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src3), b: ia3)); |
196 | src4 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src4), b: ia4)); |
197 | src1 = _mm_packus_epi32(V1: src1, V2: src2); |
198 | src3 = _mm_packus_epi32(V1: src3, V2: src4); |
199 | src1 = _mm_packus_epi16(a: src1, b: src3); |
200 | // Handle potential alpha == 0 values: |
201 | __m128i srcVectorAlphaMask = _mm_cmpeq_epi32(a: srcVectorAlpha, b: zero); |
202 | src1 = _mm_andnot_si128(a: srcVectorAlphaMask, b: src1); |
203 | // Fixup alpha values: |
204 | if (RGBx) |
205 | srcVector = _mm_or_si128(a: src1, b: alphaMask); |
206 | else |
207 | srcVector = _mm_blendv_epi8(V1: src1, V2: srcVector, M: alphaMask); |
208 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: srcVector); |
209 | } else { |
210 | if (RGBA) |
211 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: _mm_shuffle_epi8(a: srcVector, b: rgbaMask)); |
212 | else if (buffer != src) |
213 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: srcVector); |
214 | } |
215 | } else { |
216 | if (RGBx) |
217 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: alphaMask); |
218 | else |
219 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: zero); |
220 | } |
221 | } |
222 | |
223 | SIMD_EPILOGUE(i, count, 3) { |
224 | uint v = qUnpremultiply_sse4(p: src[i]); |
225 | if (RGBx) |
226 | v = 0xff000000 | v; |
227 | if (RGBA) |
228 | v = ARGB2RGBA(x: v); |
229 | buffer[i] = v; |
230 | } |
231 | } |
232 | |
233 | template<bool RGBA> |
234 | static inline void convertARGBFromRGBA64PM_sse4(uint *buffer, const QRgba64 *src, int count) |
235 | { |
236 | int i = 0; |
237 | if ((_MM_GET_EXCEPTION_MASK() & _MM_MASK_INVALID) == 0) { |
238 | for (; i < count; ++i) { |
239 | const QRgba64 v = src[i].unpremultiplied(); |
240 | buffer[i] = RGBA ? toRgba8888(rgba64: v) : toArgb32(rgba64: v); |
241 | } |
242 | return; |
243 | } |
244 | const __m128i alphaMask = _mm_set1_epi64x(q: qint64(Q_UINT64_C(0xffff) << 48)); |
245 | const __m128i alphaMask32 = _mm_set1_epi32(i: 0xff000000); |
246 | const __m128i rgbaMask = _mm_setr_epi8(b0: 2, b1: 1, b2: 0, b3: 3, b4: 6, b5: 5, b6: 4, b7: 7, b8: 10, b9: 9, b10: 8, b11: 11, b12: 14, b13: 13, b14: 12, b15: 15); |
247 | const __m128i zero = _mm_setzero_si128(); |
248 | |
249 | for (; i < count - 3; i += 4) { |
250 | __m128i srcVector1 = _mm_loadu_si128(p: (const __m128i *)&src[i]); |
251 | __m128i srcVector2 = _mm_loadu_si128(p: (const __m128i *)&src[i + 2]); |
252 | bool transparent1 = _mm_testz_si128(M: srcVector1, V: alphaMask); |
253 | bool opaque1 = _mm_testc_si128(M: srcVector1, V: alphaMask); |
254 | bool transparent2 = _mm_testz_si128(M: srcVector2, V: alphaMask); |
255 | bool opaque2 = _mm_testc_si128(M: srcVector2, V: alphaMask); |
256 | |
257 | if (!(transparent1 && transparent2)) { |
258 | if (!(opaque1 && opaque2)) { |
259 | __m128i srcVector1Alpha = _mm_srli_epi64(a: srcVector1, count: 48); |
260 | __m128i srcVector2Alpha = _mm_srli_epi64(a: srcVector2, count: 48); |
261 | __m128i srcVectorAlpha = _mm_packus_epi32(V1: srcVector1Alpha, V2: srcVector2Alpha); |
262 | const __m128 a = _mm_cvtepi32_ps(a: srcVectorAlpha); |
263 | // Convert srcVectorAlpha to final 8-bit alpha channel |
264 | srcVectorAlpha = _mm_add_epi32(a: srcVectorAlpha, b: _mm_set1_epi32(i: 128)); |
265 | srcVectorAlpha = _mm_sub_epi32(a: srcVectorAlpha, b: _mm_srli_epi32(a: srcVectorAlpha, count: 8)); |
266 | srcVectorAlpha = _mm_srli_epi32(a: srcVectorAlpha, count: 8); |
267 | srcVectorAlpha = _mm_slli_epi32(a: srcVectorAlpha, count: 24); |
268 | const __m128 ia = reciprocal_mul_ps(a, mul: 255.0f); |
269 | __m128i src1 = _mm_unpacklo_epi16(a: srcVector1, b: zero); |
270 | __m128i src2 = _mm_unpackhi_epi16(a: srcVector1, b: zero); |
271 | __m128i src3 = _mm_unpacklo_epi16(a: srcVector2, b: zero); |
272 | __m128i src4 = _mm_unpackhi_epi16(a: srcVector2, b: zero); |
273 | __m128 ia1 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(0, 0, 0, 0)); |
274 | __m128 ia2 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(1, 1, 1, 1)); |
275 | __m128 ia3 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(2, 2, 2, 2)); |
276 | __m128 ia4 = _mm_shuffle_ps(ia, ia, _MM_SHUFFLE(3, 3, 3, 3)); |
277 | src1 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src1), b: ia1)); |
278 | src2 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src2), b: ia2)); |
279 | src3 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src3), b: ia3)); |
280 | src4 = _mm_cvtps_epi32(a: _mm_mul_ps(a: _mm_cvtepi32_ps(a: src4), b: ia4)); |
281 | src1 = _mm_packus_epi32(V1: src1, V2: src2); |
282 | src3 = _mm_packus_epi32(V1: src3, V2: src4); |
283 | // Handle potential alpha == 0 values: |
284 | __m128i srcVector1AlphaMask = _mm_cmpeq_epi64(V1: srcVector1Alpha, V2: zero); |
285 | __m128i srcVector2AlphaMask = _mm_cmpeq_epi64(V1: srcVector2Alpha, V2: zero); |
286 | src1 = _mm_andnot_si128(a: srcVector1AlphaMask, b: src1); |
287 | src3 = _mm_andnot_si128(a: srcVector2AlphaMask, b: src3); |
288 | src1 = _mm_packus_epi16(a: src1, b: src3); |
289 | // Fixup alpha values: |
290 | src1 = _mm_blendv_epi8(V1: src1, V2: srcVectorAlpha, M: alphaMask32); |
291 | // Fix RGB order |
292 | if (!RGBA) |
293 | src1 = _mm_shuffle_epi8(a: src1, b: rgbaMask); |
294 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: src1); |
295 | } else { |
296 | __m128i src1 = _mm_unpacklo_epi16(a: srcVector1, b: zero); |
297 | __m128i src2 = _mm_unpackhi_epi16(a: srcVector1, b: zero); |
298 | __m128i src3 = _mm_unpacklo_epi16(a: srcVector2, b: zero); |
299 | __m128i src4 = _mm_unpackhi_epi16(a: srcVector2, b: zero); |
300 | src1 = _mm_add_epi32(a: src1, b: _mm_set1_epi32(i: 128)); |
301 | src2 = _mm_add_epi32(a: src2, b: _mm_set1_epi32(i: 128)); |
302 | src3 = _mm_add_epi32(a: src3, b: _mm_set1_epi32(i: 128)); |
303 | src4 = _mm_add_epi32(a: src4, b: _mm_set1_epi32(i: 128)); |
304 | src1 = _mm_sub_epi32(a: src1, b: _mm_srli_epi32(a: src1, count: 8)); |
305 | src2 = _mm_sub_epi32(a: src2, b: _mm_srli_epi32(a: src2, count: 8)); |
306 | src3 = _mm_sub_epi32(a: src3, b: _mm_srli_epi32(a: src3, count: 8)); |
307 | src4 = _mm_sub_epi32(a: src4, b: _mm_srli_epi32(a: src4, count: 8)); |
308 | src1 = _mm_srli_epi32(a: src1, count: 8); |
309 | src2 = _mm_srli_epi32(a: src2, count: 8); |
310 | src3 = _mm_srli_epi32(a: src3, count: 8); |
311 | src4 = _mm_srli_epi32(a: src4, count: 8); |
312 | src1 = _mm_packus_epi32(V1: src1, V2: src2); |
313 | src3 = _mm_packus_epi32(V1: src3, V2: src4); |
314 | src1 = _mm_packus_epi16(a: src1, b: src3); |
315 | if (!RGBA) |
316 | src1 = _mm_shuffle_epi8(a: src1, b: rgbaMask); |
317 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: src1); |
318 | } |
319 | } else { |
320 | _mm_storeu_si128(p: (__m128i *)&buffer[i], b: zero); |
321 | } |
322 | } |
323 | |
324 | SIMD_EPILOGUE(i, count, 3) { |
325 | buffer[i] = qConvertRgba64ToRgb32_sse4<RGBA ? PixelOrderRGB : PixelOrderBGR>(src[i]); |
326 | } |
327 | } |
328 | |
329 | #ifndef __AVX2__ |
330 | void QT_FASTCALL convertARGB32ToARGB32PM_sse4(uint *buffer, int count, const QVector<QRgb> *) |
331 | { |
332 | convertARGBToARGB32PM_sse4<false>(buffer, src: buffer, count); |
333 | } |
334 | |
335 | void QT_FASTCALL convertRGBA8888ToARGB32PM_sse4(uint *buffer, int count, const QVector<QRgb> *) |
336 | { |
337 | convertARGBToARGB32PM_sse4<true>(buffer, src: buffer, count); |
338 | } |
339 | |
340 | const QRgba64 * QT_FASTCALL convertARGB32ToRGBA64PM_sse4(QRgba64 *buffer, const uint *src, int count, |
341 | const QVector<QRgb> *, QDitherInfo *) |
342 | { |
343 | convertARGBToRGBA64PM_sse4<false>(buffer, src, count); |
344 | return buffer; |
345 | } |
346 | |
347 | const QRgba64 * QT_FASTCALL convertRGBA8888ToRGBA64PM_sse4(QRgba64 *buffer, const uint *src, int count, |
348 | const QVector<QRgb> *, QDitherInfo *) |
349 | { |
350 | convertARGBToRGBA64PM_sse4<true>(buffer, src, count); |
351 | return buffer; |
352 | } |
353 | |
354 | const uint *QT_FASTCALL fetchARGB32ToARGB32PM_sse4(uint *buffer, const uchar *src, int index, int count, |
355 | const QVector<QRgb> *, QDitherInfo *) |
356 | { |
357 | convertARGBToARGB32PM_sse4<false>(buffer, src: reinterpret_cast<const uint *>(src) + index, count); |
358 | return buffer; |
359 | } |
360 | |
361 | const uint *QT_FASTCALL fetchRGBA8888ToARGB32PM_sse4(uint *buffer, const uchar *src, int index, int count, |
362 | const QVector<QRgb> *, QDitherInfo *) |
363 | { |
364 | convertARGBToARGB32PM_sse4<true>(buffer, src: reinterpret_cast<const uint *>(src) + index, count); |
365 | return buffer; |
366 | } |
367 | |
368 | const QRgba64 *QT_FASTCALL fetchARGB32ToRGBA64PM_sse4(QRgba64 *buffer, const uchar *src, int index, int count, |
369 | const QVector<QRgb> *, QDitherInfo *) |
370 | { |
371 | convertARGBToRGBA64PM_sse4<false>(buffer, src: reinterpret_cast<const uint *>(src) + index, count); |
372 | return buffer; |
373 | } |
374 | |
375 | const QRgba64 *QT_FASTCALL fetchRGBA8888ToRGBA64PM_sse4(QRgba64 *buffer, const uchar *src, int index, int count, |
376 | const QVector<QRgb> *, QDitherInfo *) |
377 | { |
378 | convertARGBToRGBA64PM_sse4<true>(buffer, src: reinterpret_cast<const uint *>(src) + index, count); |
379 | return buffer; |
380 | } |
381 | #endif // __AVX2__ |
382 | |
383 | void QT_FASTCALL storeRGB32FromARGB32PM_sse4(uchar *dest, const uint *src, int index, int count, |
384 | const QVector<QRgb> *, QDitherInfo *) |
385 | { |
386 | uint *d = reinterpret_cast<uint *>(dest) + index; |
387 | convertARGBFromARGB32PM_sse4<false,true>(buffer: d, src, count); |
388 | } |
389 | |
390 | void QT_FASTCALL storeARGB32FromARGB32PM_sse4(uchar *dest, const uint *src, int index, int count, |
391 | const QVector<QRgb> *, QDitherInfo *) |
392 | { |
393 | uint *d = reinterpret_cast<uint *>(dest) + index; |
394 | convertARGBFromARGB32PM_sse4<false,false>(buffer: d, src, count); |
395 | } |
396 | |
397 | void QT_FASTCALL storeRGBA8888FromARGB32PM_sse4(uchar *dest, const uint *src, int index, int count, |
398 | const QVector<QRgb> *, QDitherInfo *) |
399 | { |
400 | uint *d = reinterpret_cast<uint *>(dest) + index; |
401 | convertARGBFromARGB32PM_sse4<true,false>(buffer: d, src, count); |
402 | } |
403 | |
404 | void QT_FASTCALL storeRGBXFromARGB32PM_sse4(uchar *dest, const uint *src, int index, int count, |
405 | const QVector<QRgb> *, QDitherInfo *) |
406 | { |
407 | uint *d = reinterpret_cast<uint *>(dest) + index; |
408 | convertARGBFromARGB32PM_sse4<true,true>(buffer: d, src, count); |
409 | } |
410 | |
411 | template<QtPixelOrder PixelOrder> |
412 | void QT_FASTCALL storeA2RGB30PMFromARGB32PM_sse4(uchar *dest, const uint *src, int index, int count, |
413 | const QVector<QRgb> *, QDitherInfo *) |
414 | { |
415 | uint *d = reinterpret_cast<uint *>(dest) + index; |
416 | for (int i = 0; i < count; ++i) |
417 | d[i] = qConvertArgb32ToA2rgb30_sse4<PixelOrder>(src[i]); |
418 | } |
419 | |
420 | #if QT_CONFIG(raster_64bit) |
421 | void QT_FASTCALL destStore64ARGB32_sse4(QRasterBuffer *rasterBuffer, int x, int y, const QRgba64 *buffer, int length) |
422 | { |
423 | uint *dest = (uint*)rasterBuffer->scanLine(y) + x; |
424 | convertARGBFromRGBA64PM_sse4<false>(buffer: dest, src: buffer, count: length); |
425 | } |
426 | |
427 | void QT_FASTCALL destStore64RGBA8888_sse4(QRasterBuffer *rasterBuffer, int x, int y, const QRgba64 *buffer, int length) |
428 | { |
429 | uint *dest = (uint*)rasterBuffer->scanLine(y) + x; |
430 | convertARGBFromRGBA64PM_sse4<true>(buffer: dest, src: buffer, count: length); |
431 | } |
432 | #endif |
433 | |
434 | void QT_FASTCALL storeARGB32FromRGBA64PM_sse4(uchar *dest, const QRgba64 *src, int index, int count, |
435 | const QVector<QRgb> *, QDitherInfo *) |
436 | { |
437 | uint *d = (uint*)dest + index; |
438 | convertARGBFromRGBA64PM_sse4<false>(buffer: d, src, count); |
439 | } |
440 | |
441 | void QT_FASTCALL storeRGBA8888FromRGBA64PM_sse4(uchar *dest, const QRgba64 *src, int index, int count, |
442 | const QVector<QRgb> *, QDitherInfo *) |
443 | { |
444 | uint *d = (uint*)dest + index; |
445 | convertARGBFromRGBA64PM_sse4<true>(buffer: d, src, count); |
446 | } |
447 | |
448 | template |
449 | void QT_FASTCALL storeA2RGB30PMFromARGB32PM_sse4<PixelOrderBGR>(uchar *dest, const uint *src, int index, int count, |
450 | const QVector<QRgb> *, QDitherInfo *); |
451 | template |
452 | void QT_FASTCALL storeA2RGB30PMFromARGB32PM_sse4<PixelOrderRGB>(uchar *dest, const uint *src, int index, int count, |
453 | const QVector<QRgb> *, QDitherInfo *); |
454 | |
455 | QT_END_NAMESPACE |
456 | |
457 | #endif |
458 | |