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
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
|
// Copyright 2023 Intel Corporation
// SPDX-License-Identifier: Apache-2.0
#pragma once
#include "common/OSPCommon.ih"
#include "rkcommon/math/vec.ih"
OSPRAY_BEGIN_ISPC_NAMESPACE
// helpers
//////////////////////////////////////////////////////////////////////////////
#define __foreach_fetcher(FCT) \
FCT(RGBA8); \
FCT(SRGBA); \
FCT(RGBA32F); \
FCT(RGB8); \
FCT(SRGB); \
FCT(LA8); \
FCT(RA8); \
FCT(RGB32F); \
FCT(RA32F); \
FCT(R8); \
FCT(L8); \
FCT(R32F); \
FCT(RGBA16); \
FCT(RGB16); \
FCT(RA16); \
FCT(R16); \
FCT(RGBA16F); \
FCT(RGB16F); \
FCT(RA16F); \
FCT(R16F)
#define BYTE2FLOAT (1.f / 255.f)
#define SHORT2FLOAT (1.f / 65535.f)
#ifdef OSPRAY_TARGET_SYCL
using half = sycl::half;
inline float half_to_float(int16 x)
{
return *reinterpret_cast<half *>(&x);
}
#else
typedef int16 half;
// Low-level texel accessors: setters
//////////////////////////////////////////////////////////////////////////////
inline void setTexel_RGBA8(
void *uniform data, const uint32 idx, const vec4f value)
{
((uint32 * uniform) data)[idx] = cvt_uint32(value);
}
inline void setTexel_RGBA16(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 texelOfs = idx * 2; // write as 2x 32bit
const uint32 r = cvt_uint16(value.x);
const uint32 g = cvt_uint16(value.y);
((uint32 * uniform) data)[texelOfs] = r | (g << 16);
const uint32 b = cvt_uint16(value.z);
const uint32 a = cvt_uint16(value.w);
((uint32 * uniform) data)[texelOfs + 1] = b | (a << 16);
}
inline void setTexel_RGB8(
void *uniform data, const uint32 idx, const vec4f value)
{
uint8 *uniform texel = (uint8 * uniform) data;
const uint32 texelOfs = 3 * idx;
texel[texelOfs] = cvt_uint32(value.x);
texel[texelOfs + 1] = cvt_uint32(value.y);
texel[texelOfs + 2] = cvt_uint32(value.z);
}
inline void setTexel_RGB16(
void *uniform data, const uint32 idx, const vec4f value)
{
uint16 *uniform texel = (uint16 * uniform) data;
const uint32 texelOfs = 3 * idx;
texel[texelOfs] = cvt_uint16(value.x);
texel[texelOfs + 1] = cvt_uint16(value.y);
texel[texelOfs + 2] = cvt_uint16(value.z);
}
inline void setTexel_RA8(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 r = cvt_uint32(value.x);
const uint32 a = cvt_uint32(value.w);
uint32 c = r | (a << 8);
((uint16 * uniform) data)[idx] = c;
}
inline void setTexel_RA16(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 r = cvt_uint16(value.x);
const uint32 a = cvt_uint16(value.w);
uint32 c = r | (a << 16);
((uint32 * uniform) data)[idx] = c;
}
inline void setTexel_R8(void *uniform data, const uint32 idx, const vec4f value)
{
((uint8 * uniform) data)[idx] = cvt_uint32(value.x);
}
inline void setTexel_R16(
void *uniform data, const uint32 idx, const vec4f value)
{
((uint16 * uniform) data)[idx] = cvt_uint16(value.x);
}
// luminance (basically gamma-corrected grayscale) with alpha
inline void setTexel_LA8(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 l = cvt_uint32(linear_to_srgb(value.x));
const uint32 a = cvt_uint32(value.w);
uint32 c = l | a << 8;
((uint16 * uniform) data)[idx] = c;
}
// luminance (basically gamma-corrected grayscale)
inline void setTexel_L8(void *uniform data, const uint32 idx, const vec4f value)
{
((uint8 * uniform) data)[idx] = cvt_uint32(linear_to_srgb(value.x));
}
inline void setTexel_SRGBA(
void *uniform data, const uint32 idx, const vec4f value)
{
((uint32 * uniform) data)[idx] = linear_to_srgba8(value);
}
inline void setTexel_SRGB(
void *uniform data, const uint32 idx, const vec4f value)
{
uint8 *uniform texel = (uint8 * uniform) data;
const uint32 texelOfs = 3 * idx;
const vec4f v = linear_to_srgba(value);
texel[texelOfs] = cvt_uint32(v.x);
texel[texelOfs + 1] = cvt_uint32(v.y);
texel[texelOfs + 2] = cvt_uint32(v.z);
}
inline void setTexel_RGBA32F(
void *uniform data, const uint32 idx, const vec4f value)
{
((vec4f * uniform) data)[idx] = value;
}
inline void setTexel_RGB32F(
void *uniform data, const uint32 idx, const vec4f value)
{
((vec3f * uniform) data)[idx] = make_vec3f(value.x, value.y, value.z);
}
inline void setTexel_RA32F(
void *uniform data, const uint32 idx, const vec4f value)
{
((vec2f * uniform) data)[idx] = make_vec2f(value.x, value.w);
}
inline void setTexel_R32F(
void *uniform data, const uint32 idx, const vec4f value)
{
((float *uniform)data)[idx] = value.x;
}
inline vec4f setTexel_RGBA16F(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 texelOfs = idx * 2; // write as 2x 32bit
const uint32 r = float_to_half(value.x);
const uint32 g = float_to_half(value.y);
((uint32 * uniform) data)[texelOfs] = r | (g << 16);
const uint32 b = float_to_half(value.z);
const uint32 a = float_to_half(value.w);
((uint32 * uniform) data)[texelOfs + 1] = b | (a << 16);
}
inline vec4f setTexel_RGB16F(
void *uniform data, const uint32 idx, const vec4f value)
{
half *uniform texel = (half * uniform) data;
const uint32 texelOfs = 3 * idx;
texel[texelOfs] = float_to_half(value.x);
texel[texelOfs + 1] = float_to_half(value.y);
texel[texelOfs + 2] = float_to_half(value.z);
}
inline vec4f setTexel_RA16F(
void *uniform data, const uint32 idx, const vec4f value)
{
const uint32 r = float_to_half(value.x);
const uint32 a = float_to_half(value.w);
((uint32 * uniform) data)[idx] = r | (a << 16);
}
inline vec4f setTexel_R16F(
void *uniform data, const uint32 idx, const vec4f value)
{
((half * uniform) data)[idx] = float_to_half(value.x);
}
#endif
// Low-level texel accessors: getters
//////////////////////////////////////////////////////////////////////////////
inline vec4f getTexel_RGBA8(const void *data, const uint32 idx)
{
assert(data);
const uint32 c = ((const uint32 *)data)[idx];
const uint32 r = c & 0xff;
const uint32 g = (c >> 8) & 0xff;
const uint32 b = (c >> 16) & 0xff;
const uint32 a = c >> 24;
return make_vec4f((float)r, (float)g, (float)b, (float)a) * BYTE2FLOAT;
}
inline vec4f getTexel_RGBA16(const void *data, const uint32 idx)
{
assert(data);
const uint32 texelOfs = idx * 2; // fetch as 2x 32bit
uint32 c = ((const uint32 *)data)[texelOfs];
const uint32 r = c & 0xffff;
const uint32 g = c >> 16;
c = ((const uint32 *)data)[texelOfs + 1];
const uint32 b = c & 0xffff;
const uint32 a = c >> 16;
return make_vec4f((float)r, (float)g, (float)b, (float)a) * SHORT2FLOAT;
}
inline vec4f getTexel_RGB8(const void *data, const uint32 idx)
{
assert(data);
const uint8 *texel = (const uint8 *)data;
const uint32 texelOfs = 3 * idx;
const uint32 r = texel[texelOfs];
const uint32 g = texel[texelOfs + 1];
const uint32 b = texel[texelOfs + 2];
return make_vec4f(make_vec3f((float)r, (float)g, (float)b) * BYTE2FLOAT, 1.f);
}
inline vec4f getTexel_RGB16(const void *data, const uint32 idx)
{
assert(data);
const uint16 *texel = (const uint16 *)data;
const uint32 texelOfs = 3 * idx;
const uint32 r = texel[texelOfs];
const uint32 g = texel[texelOfs + 1];
const uint32 b = texel[texelOfs + 2];
return make_vec4f(
make_vec3f((float)r, (float)g, (float)b) * SHORT2FLOAT, 1.f);
}
inline vec4f getTexel_RA8(const void *data, const uint32 idx)
{
assert(data);
const uint32 t = ((const uint16 *)data)[idx];
const uint32 c = t & 0xff;
const uint32 a = t >> 8;
return make_vec4f(c * BYTE2FLOAT, 0.f, 0.f, a * BYTE2FLOAT);
}
inline vec4f getTexel_RA16(const void *data, const uint32 idx)
{
assert(data);
const uint32 t = ((const uint32 *)data)[idx];
const uint32 c = t & 0xffff;
const uint32 a = t >> 16;
return make_vec4f(c * SHORT2FLOAT, 0.f, 0.f, a * SHORT2FLOAT);
}
inline vec4f getTexel_R8(const void *data, const uint32 idx)
{
assert(data);
const uint8 c = ((const uint8 *)data)[idx];
return make_vec4f(c * BYTE2FLOAT, 0.f, 0.f, 1.f);
}
inline vec4f getTexel_R16(const void *data, const uint32 idx)
{
assert(data);
const uint16 c = ((const uint16 *)data)[idx];
return make_vec4f(c * SHORT2FLOAT, 0.f, 0.f, 1.f);
}
// luminance (basically gamma-corrected grayscale) with alpha
inline vec4f getTexel_LA8(const void *data, const uint32 idx)
{
assert(data);
const uint32 t = ((const uint16 *)data)[idx];
const uint32 l = t & 0xff;
const uint32 a = t >> 8;
return make_vec4f(make_vec3f(srgb_to_linear(l * BYTE2FLOAT)), a * BYTE2FLOAT);
}
// luminance (basically gamma-corrected grayscale)
inline vec4f getTexel_L8(const void *data, const uint32 idx)
{
assert(data);
const uint8 l = ((const uint8 *)data)[idx];
return make_vec4f(make_vec3f(srgb_to_linear(l * BYTE2FLOAT)), 1.f);
}
inline vec4f getTexel_SRGBA(const void *data, const uint32 idx)
{
return srgba_to_linear(getTexel_RGBA8(data, idx));
}
inline vec4f getTexel_SRGB(const void *data, const uint32 idx)
{
return srgba_to_linear(getTexel_RGB8(data, idx));
}
inline vec4f getTexel_RGBA32F(const void *data, const uint32 idx)
{
assert(data);
return ((const vec4f *)data)[idx];
}
inline vec4f getTexel_RGB32F(const void *data, const uint32 idx)
{
assert(data);
vec3f v = ((const vec3f *)data)[idx];
return make_vec4f(v, 1.f);
}
inline vec4f getTexel_RA32F(const void *data, const uint32 idx)
{
assert(data);
vec2f v = ((const vec2f *)data)[idx];
return make_vec4f(v.x, 0.0f, 0.0f, v.y);
}
inline vec4f getTexel_R32F(const void *data, const uint32 idx)
{
assert(data);
float v = ((const float *)data)[idx];
return make_vec4f(v, 0.f, 0.f, 1.f);
}
inline vec4f getTexel_RGBA16F(const void *data, const uint32 idx)
{
assert(data);
const uint32 texelOfs = idx * 2; // fetch as 2x 32bit
uint32 c = ((const uint32 *)data)[texelOfs];
const int16 r = c & 0xffff;
const int16 g = c >> 16;
c = ((const uint32 *)data)[texelOfs + 1];
const int16 b = c & 0xffff;
const int16 a = c >> 16;
return make_vec4f(
half_to_float(r), half_to_float(g), half_to_float(b), half_to_float(a));
}
inline vec4f getTexel_RGB16F(const void *data, const uint32 idx)
{
assert(data);
const half *texel = (const half *)data;
const uint32 texelOfs = 3 * idx;
const half r = texel[texelOfs];
const half g = texel[texelOfs + 1];
const half b = texel[texelOfs + 2];
return make_vec4f(half_to_float(r), half_to_float(g), half_to_float(b), 1.f);
}
inline vec4f getTexel_RA16F(const void *data, const uint32 idx)
{
assert(data);
const uint32 t = ((const uint32 *)data)[idx];
const int16 c = t & 0xffff;
const int16 a = t >> 16;
return make_vec4f(half_to_float(c), 0.0f, 0.0f, half_to_float(a));
}
inline vec4f getTexel_R16F(const void *data, const uint32 idx)
{
assert(data);
const half c = ((const half *)data)[idx];
return make_vec4f(half_to_float(c), 0.f, 0.f, 1.f);
}
#undef BYTE2FLOAT
#undef SHORT2FLOAT
OSPRAY_END_ISPC_NAMESPACE
|
