1 #include "rs_types.rsh" 2 3 // Conversions 4 #define CVT_FUNC_2(typeout, typein) \ 5 extern typeout##2 __attribute__((overloadable)) \ 6 convert_##typeout##2(typein##2 v) { \ 7 typeout##2 r = {(typeout)v.x, (typeout)v.y}; \ 8 return r; \ 9 } \ 10 extern typeout##3 __attribute__((overloadable)) \ 11 convert_##typeout##3(typein##3 v) { \ 12 typeout##3 r = {(typeout)v.x, (typeout)v.y, (typeout)v.z}; \ 13 return r; \ 14 } \ 15 extern typeout##4 __attribute__((overloadable)) \ 16 convert_##typeout##4(typein##4 v) { \ 17 typeout##4 r = {(typeout)v.x, (typeout)v.y, (typeout)v.z, \ 18 (typeout)v.w}; \ 19 return r; \ 20 } 21 22 #define CVT_FUNC(type) CVT_FUNC_2(type, uchar) \ 23 CVT_FUNC_2(type, char) \ 24 CVT_FUNC_2(type, ushort) \ 25 CVT_FUNC_2(type, short) \ 26 CVT_FUNC_2(type, uint) \ 27 CVT_FUNC_2(type, int) \ 28 CVT_FUNC_2(type, float) 29 30 CVT_FUNC(char) 31 //CVT_FUNC(uchar) 32 CVT_FUNC(short) 33 CVT_FUNC(ushort) 34 CVT_FUNC(int) 35 CVT_FUNC(uint) 36 //CVT_FUNC(float) 37 38 // Float ops, 6.11.2 39 40 #define FN_FUNC_FN(fnc) \ 41 extern float2 __attribute__((overloadable)) fnc(float2 v) { \ 42 float2 r; \ 43 r.x = fnc(v.x); \ 44 r.y = fnc(v.y); \ 45 return r; \ 46 } \ 47 extern float3 __attribute__((overloadable)) fnc(float3 v) { \ 48 float3 r; \ 49 r.x = fnc(v.x); \ 50 r.y = fnc(v.y); \ 51 r.z = fnc(v.z); \ 52 return r; \ 53 } \ 54 extern float4 __attribute__((overloadable)) fnc(float4 v) { \ 55 float4 r; \ 56 r.x = fnc(v.x); \ 57 r.y = fnc(v.y); \ 58 r.z = fnc(v.z); \ 59 r.w = fnc(v.w); \ 60 return r; \ 61 } 62 63 #define IN_FUNC_FN(fnc) \ 64 extern int2 __attribute__((overloadable)) fnc(float2 v) { \ 65 int2 r; \ 66 r.x = fnc(v.x); \ 67 r.y = fnc(v.y); \ 68 return r; \ 69 } \ 70 extern int3 __attribute__((overloadable)) fnc(float3 v) { \ 71 int3 r; \ 72 r.x = fnc(v.x); \ 73 r.y = fnc(v.y); \ 74 r.z = fnc(v.z); \ 75 return r; \ 76 } \ 77 extern int4 __attribute__((overloadable)) fnc(float4 v) { \ 78 int4 r; \ 79 r.x = fnc(v.x); \ 80 r.y = fnc(v.y); \ 81 r.z = fnc(v.z); \ 82 r.w = fnc(v.w); \ 83 return r; \ 84 } 85 86 #define FN_FUNC_FN_FN(fnc) \ 87 extern float2 __attribute__((overloadable)) fnc(float2 v1, float2 v2) { \ 88 float2 r; \ 89 r.x = fnc(v1.x, v2.x); \ 90 r.y = fnc(v1.y, v2.y); \ 91 return r; \ 92 } \ 93 extern float3 __attribute__((overloadable)) fnc(float3 v1, float3 v2) { \ 94 float3 r; \ 95 r.x = fnc(v1.x, v2.x); \ 96 r.y = fnc(v1.y, v2.y); \ 97 r.z = fnc(v1.z, v2.z); \ 98 return r; \ 99 } \ 100 extern float4 __attribute__((overloadable)) fnc(float4 v1, float4 v2) { \ 101 float4 r; \ 102 r.x = fnc(v1.x, v2.x); \ 103 r.y = fnc(v1.y, v2.y); \ 104 r.z = fnc(v1.z, v2.z); \ 105 r.w = fnc(v1.w, v2.w); \ 106 return r; \ 107 } 108 109 #define FN_FUNC_FN_F(fnc) \ 110 extern float2 __attribute__((overloadable)) fnc(float2 v1, float v2) { \ 111 float2 r; \ 112 r.x = fnc(v1.x, v2); \ 113 r.y = fnc(v1.y, v2); \ 114 return r; \ 115 } \ 116 extern float3 __attribute__((overloadable)) fnc(float3 v1, float v2) { \ 117 float3 r; \ 118 r.x = fnc(v1.x, v2); \ 119 r.y = fnc(v1.y, v2); \ 120 r.z = fnc(v1.z, v2); \ 121 return r; \ 122 } \ 123 extern float4 __attribute__((overloadable)) fnc(float4 v1, float v2) { \ 124 float4 r; \ 125 r.x = fnc(v1.x, v2); \ 126 r.y = fnc(v1.y, v2); \ 127 r.z = fnc(v1.z, v2); \ 128 r.w = fnc(v1.w, v2); \ 129 return r; \ 130 } 131 132 #define FN_FUNC_FN_IN(fnc) \ 133 extern float2 __attribute__((overloadable)) fnc(float2 v1, int2 v2) { \ 134 float2 r; \ 135 r.x = fnc(v1.x, v2.x); \ 136 r.y = fnc(v1.y, v2.y); \ 137 return r; \ 138 } \ 139 extern float3 __attribute__((overloadable)) fnc(float3 v1, int3 v2) { \ 140 float3 r; \ 141 r.x = fnc(v1.x, v2.x); \ 142 r.y = fnc(v1.y, v2.y); \ 143 r.z = fnc(v1.z, v2.z); \ 144 return r; \ 145 } \ 146 extern float4 __attribute__((overloadable)) fnc(float4 v1, int4 v2) { \ 147 float4 r; \ 148 r.x = fnc(v1.x, v2.x); \ 149 r.y = fnc(v1.y, v2.y); \ 150 r.z = fnc(v1.z, v2.z); \ 151 r.w = fnc(v1.w, v2.w); \ 152 return r; \ 153 } 154 155 #define FN_FUNC_FN_I(fnc) \ 156 extern float2 __attribute__((overloadable)) fnc(float2 v1, int v2) { \ 157 float2 r; \ 158 r.x = fnc(v1.x, v2); \ 159 r.y = fnc(v1.y, v2); \ 160 return r; \ 161 } \ 162 extern float3 __attribute__((overloadable)) fnc(float3 v1, int v2) { \ 163 float3 r; \ 164 r.x = fnc(v1.x, v2); \ 165 r.y = fnc(v1.y, v2); \ 166 r.z = fnc(v1.z, v2); \ 167 return r; \ 168 } \ 169 extern float4 __attribute__((overloadable)) fnc(float4 v1, int v2) { \ 170 float4 r; \ 171 r.x = fnc(v1.x, v2); \ 172 r.y = fnc(v1.y, v2); \ 173 r.z = fnc(v1.z, v2); \ 174 r.w = fnc(v1.w, v2); \ 175 return r; \ 176 } 177 178 #define FN_FUNC_FN_PFN(fnc) \ 179 extern float2 __attribute__((overloadable)) \ 180 fnc(float2 v1, float2 *v2) { \ 181 float2 r; \ 182 float t[2]; \ 183 r.x = fnc(v1.x, &t[0]); \ 184 r.y = fnc(v1.y, &t[1]); \ 185 v2->x = t[0]; \ 186 v2->y = t[1]; \ 187 return r; \ 188 } \ 189 extern float3 __attribute__((overloadable)) \ 190 fnc(float3 v1, float3 *v2) { \ 191 float3 r; \ 192 float t[3]; \ 193 r.x = fnc(v1.x, &t[0]); \ 194 r.y = fnc(v1.y, &t[1]); \ 195 r.z = fnc(v1.z, &t[2]); \ 196 v2->x = t[0]; \ 197 v2->y = t[1]; \ 198 v2->z = t[2]; \ 199 return r; \ 200 } \ 201 extern float4 __attribute__((overloadable)) \ 202 fnc(float4 v1, float4 *v2) { \ 203 float4 r; \ 204 float t[4]; \ 205 r.x = fnc(v1.x, &t[0]); \ 206 r.y = fnc(v1.y, &t[1]); \ 207 r.z = fnc(v1.z, &t[2]); \ 208 r.w = fnc(v1.w, &t[3]); \ 209 v2->x = t[0]; \ 210 v2->y = t[1]; \ 211 v2->z = t[2]; \ 212 v2->w = t[3]; \ 213 return r; \ 214 } 215 216 #define FN_FUNC_FN_PIN(fnc) \ 217 extern float2 __attribute__((overloadable)) fnc(float2 v1, int2 *v2) { \ 218 float2 r; \ 219 int t[2]; \ 220 r.x = fnc(v1.x, &t[0]); \ 221 r.y = fnc(v1.y, &t[1]); \ 222 v2->x = t[0]; \ 223 v2->y = t[1]; \ 224 return r; \ 225 } \ 226 extern float3 __attribute__((overloadable)) fnc(float3 v1, int3 *v2) { \ 227 float3 r; \ 228 int t[3]; \ 229 r.x = fnc(v1.x, &t[0]); \ 230 r.y = fnc(v1.y, &t[1]); \ 231 r.z = fnc(v1.z, &t[2]); \ 232 v2->x = t[0]; \ 233 v2->y = t[1]; \ 234 v2->z = t[2]; \ 235 return r; \ 236 } \ 237 extern float4 __attribute__((overloadable)) fnc(float4 v1, int4 *v2) { \ 238 float4 r; \ 239 int t[4]; \ 240 r.x = fnc(v1.x, &t[0]); \ 241 r.y = fnc(v1.y, &t[1]); \ 242 r.z = fnc(v1.z, &t[2]); \ 243 r.w = fnc(v1.w, &t[3]); \ 244 v2->x = t[0]; \ 245 v2->y = t[1]; \ 246 v2->z = t[2]; \ 247 v2->w = t[3]; \ 248 return r; \ 249 } 250 251 #define FN_FUNC_FN_FN_FN(fnc) \ 252 extern float2 __attribute__((overloadable)) \ 253 fnc(float2 v1, float2 v2, float2 v3) { \ 254 float2 r; \ 255 r.x = fnc(v1.x, v2.x, v3.x); \ 256 r.y = fnc(v1.y, v2.y, v3.y); \ 257 return r; \ 258 } \ 259 extern float3 __attribute__((overloadable)) \ 260 fnc(float3 v1, float3 v2, float3 v3) { \ 261 float3 r; \ 262 r.x = fnc(v1.x, v2.x, v3.x); \ 263 r.y = fnc(v1.y, v2.y, v3.y); \ 264 r.z = fnc(v1.z, v2.z, v3.z); \ 265 return r; \ 266 } \ 267 extern float4 __attribute__((overloadable)) \ 268 fnc(float4 v1, float4 v2, float4 v3) { \ 269 float4 r; \ 270 r.x = fnc(v1.x, v2.x, v3.x); \ 271 r.y = fnc(v1.y, v2.y, v3.y); \ 272 r.z = fnc(v1.z, v2.z, v3.z); \ 273 r.w = fnc(v1.w, v2.w, v3.w); \ 274 return r; \ 275 } 276 277 #define FN_FUNC_FN_FN_PIN(fnc) \ 278 extern float2 __attribute__((overloadable)) \ 279 fnc(float2 v1, float2 v2, int2 *v3) { \ 280 float2 r; \ 281 int t[2]; \ 282 r.x = fnc(v1.x, v2.x, &t[0]); \ 283 r.y = fnc(v1.y, v2.y, &t[1]); \ 284 v3->x = t[0]; \ 285 v3->y = t[1]; \ 286 return r; \ 287 } \ 288 extern float3 __attribute__((overloadable)) \ 289 fnc(float3 v1, float3 v2, int3 *v3) { \ 290 float3 r; \ 291 int t[3]; \ 292 r.x = fnc(v1.x, v2.x, &t[0]); \ 293 r.y = fnc(v1.y, v2.y, &t[1]); \ 294 r.z = fnc(v1.z, v2.z, &t[2]); \ 295 v3->x = t[0]; \ 296 v3->y = t[1]; \ 297 v3->z = t[2]; \ 298 return r; \ 299 } \ 300 extern float4 __attribute__((overloadable)) \ 301 fnc(float4 v1, float4 v2, int4 *v3) { \ 302 float4 r; \ 303 int t[4]; \ 304 r.x = fnc(v1.x, v2.x, &t[0]); \ 305 r.y = fnc(v1.y, v2.y, &t[1]); \ 306 r.z = fnc(v1.z, v2.z, &t[2]); \ 307 r.w = fnc(v1.w, v2.w, &t[3]); \ 308 v3->x = t[0]; \ 309 v3->y = t[1]; \ 310 v3->z = t[2]; \ 311 v3->w = t[3]; \ 312 return r; \ 313 } 314 315 static const int iposinf = 0x7f800000; 316 static const int ineginf = 0xff800000; 317 318 static const float posinf() { 319 float f = *((float*)&iposinf); 320 return f; 321 } 322 323 static const float neginf() { 324 float f = *((float*)&ineginf); 325 return f; 326 } 327 328 static bool isinf(float f) { 329 int i = *((int*)(void*)&f); 330 return (i == iposinf) || (i == ineginf); 331 } 332 333 static bool isnan(float f) { 334 int i = *((int*)(void*)&f); 335 return (((i & 0x7f800000) == 0x7f800000) && (i & 0x007fffff)); 336 } 337 338 static bool isposzero(float f) { 339 int i = *((int*)(void*)&f); 340 return (i == 0x00000000); 341 } 342 343 static bool isnegzero(float f) { 344 int i = *((int*)(void*)&f); 345 return (i == 0x80000000); 346 } 347 348 static bool iszero(float f) { 349 return isposzero(f) || isnegzero(f); 350 } 351 352 353 extern float __attribute__((overloadable)) acos(float); 354 FN_FUNC_FN(acos) 355 356 extern float __attribute__((overloadable)) acosh(float); 357 FN_FUNC_FN(acosh) 358 359 360 extern float __attribute__((overloadable)) acospi(float v) { 361 return acos(v) / M_PI; 362 } 363 FN_FUNC_FN(acospi) 364 365 extern float __attribute__((overloadable)) asin(float); 366 FN_FUNC_FN(asin) 367 368 extern float __attribute__((overloadable)) asinh(float); 369 FN_FUNC_FN(asinh) 370 371 extern float __attribute__((overloadable)) asinpi(float v) { 372 return asin(v) / M_PI; 373 } 374 FN_FUNC_FN(asinpi) 375 376 extern float __attribute__((overloadable)) atan(float); 377 FN_FUNC_FN(atan) 378 379 extern float __attribute__((overloadable)) atan2(float, float); 380 FN_FUNC_FN_FN(atan2) 381 382 extern float __attribute__((overloadable)) atanh(float); 383 FN_FUNC_FN(atanh) 384 385 extern float __attribute__((overloadable)) atanpi(float v) { 386 return atan(v) / M_PI; 387 } 388 FN_FUNC_FN(atanpi) 389 390 391 extern float __attribute__((overloadable)) atan2pi(float y, float x) { 392 return atan2(y, x) / M_PI; 393 } 394 FN_FUNC_FN_FN(atan2pi) 395 396 extern float __attribute__((overloadable)) cbrt(float); 397 FN_FUNC_FN(cbrt) 398 399 extern float __attribute__((overloadable)) ceil(float); 400 FN_FUNC_FN(ceil) 401 402 extern float __attribute__((overloadable)) copysign(float, float); 403 FN_FUNC_FN_FN(copysign) 404 405 extern float __attribute__((overloadable)) cos(float); 406 FN_FUNC_FN(cos) 407 408 extern float __attribute__((overloadable)) cosh(float); 409 FN_FUNC_FN(cosh) 410 411 extern float __attribute__((overloadable)) cospi(float v) { 412 return cos(v * M_PI); 413 } 414 FN_FUNC_FN(cospi) 415 416 extern float __attribute__((overloadable)) erfc(float); 417 FN_FUNC_FN(erfc) 418 419 extern float __attribute__((overloadable)) erf(float); 420 FN_FUNC_FN(erf) 421 422 extern float __attribute__((overloadable)) exp(float); 423 FN_FUNC_FN(exp) 424 425 extern float __attribute__((overloadable)) exp2(float); 426 FN_FUNC_FN(exp2) 427 428 extern float __attribute__((overloadable)) pow(float, float); 429 430 extern float __attribute__((overloadable)) exp10(float v) { 431 return pow(10.f, v); 432 } 433 FN_FUNC_FN(exp10) 434 435 extern float __attribute__((overloadable)) expm1(float); 436 FN_FUNC_FN(expm1) 437 438 extern float __attribute__((overloadable)) fabs(float); 439 FN_FUNC_FN(fabs) 440 441 extern float __attribute__((overloadable)) fdim(float, float); 442 FN_FUNC_FN_FN(fdim) 443 444 extern float __attribute__((overloadable)) floor(float); 445 FN_FUNC_FN(floor) 446 447 extern float __attribute__((overloadable)) fma(float, float, float); 448 FN_FUNC_FN_FN_FN(fma) 449 450 extern float __attribute__((overloadable)) fmax(float, float); 451 FN_FUNC_FN_FN(fmax); 452 FN_FUNC_FN_F(fmax); 453 454 extern float __attribute__((overloadable)) fmin(float, float); 455 FN_FUNC_FN_FN(fmin); 456 FN_FUNC_FN_F(fmin); 457 458 extern float __attribute__((overloadable)) fmod(float, float); 459 FN_FUNC_FN_FN(fmod) 460 461 extern float __attribute__((overloadable)) fract(float v, float *iptr) { 462 int i = (int)floor(v); 463 iptr[0] = i; 464 return fmin(v - i, 0x1.fffffep-1f); 465 } 466 FN_FUNC_FN_PFN(fract) 467 468 extern float __attribute__((overloadable)) frexp(float, int *); 469 FN_FUNC_FN_PIN(frexp) 470 471 extern float __attribute__((overloadable)) hypot(float, float); 472 FN_FUNC_FN_FN(hypot) 473 474 extern int __attribute__((overloadable)) ilogb(float); 475 IN_FUNC_FN(ilogb) 476 477 extern float __attribute__((overloadable)) ldexp(float, int); 478 FN_FUNC_FN_IN(ldexp) 479 FN_FUNC_FN_I(ldexp) 480 481 extern float __attribute__((overloadable)) lgamma(float); 482 FN_FUNC_FN(lgamma) 483 extern float __attribute__((overloadable)) lgamma(float, int*); 484 FN_FUNC_FN_PIN(lgamma) 485 486 extern float __attribute__((overloadable)) log(float); 487 FN_FUNC_FN(log) 488 489 extern float __attribute__((overloadable)) log10(float); 490 FN_FUNC_FN(log10) 491 492 493 extern float __attribute__((overloadable)) log2(float v) { 494 return log10(v) / log10(2.f); 495 } 496 FN_FUNC_FN(log2) 497 498 extern float __attribute__((overloadable)) log1p(float); 499 FN_FUNC_FN(log1p) 500 501 extern float __attribute__((overloadable)) logb(float); 502 FN_FUNC_FN(logb) 503 504 extern float __attribute__((overloadable)) mad(float a, float b, float c) { 505 return a * b + c; 506 } 507 extern float2 __attribute__((overloadable)) mad(float2 a, float2 b, float2 c) { 508 return a * b + c; 509 } 510 extern float3 __attribute__((overloadable)) mad(float3 a, float3 b, float3 c) { 511 return a * b + c; 512 } 513 extern float4 __attribute__((overloadable)) mad(float4 a, float4 b, float4 c) { 514 return a * b + c; 515 } 516 517 extern float __attribute__((overloadable)) modf(float, float *); 518 FN_FUNC_FN_PFN(modf); 519 520 extern float __attribute__((overloadable)) nan(uint v) { 521 float f[1]; 522 uint32_t *ip = (uint32_t *)f; 523 *ip = v | 0x7fc00000; 524 return f[0]; 525 } 526 527 extern float __attribute__((overloadable)) nextafter(float, float); 528 FN_FUNC_FN_FN(nextafter) 529 530 FN_FUNC_FN_FN(pow) 531 532 extern float __attribute__((overloadable)) pown(float v, int p) { 533 return pow(v, (float)p); 534 } 535 extern float2 __attribute__((overloadable)) pown(float2 v, int2 p) { 536 return pow(v, (float2)p); 537 } 538 extern float3 __attribute__((overloadable)) pown(float3 v, int3 p) { 539 return pow(v, (float3)p); 540 } 541 extern float4 __attribute__((overloadable)) pown(float4 v, int4 p) { 542 return pow(v, (float4)p); 543 } 544 545 extern float __attribute__((overloadable)) powr(float v, float p) { 546 return pow(v, p); 547 } 548 extern float2 __attribute__((overloadable)) powr(float2 v, float2 p) { 549 return pow(v, p); 550 } 551 extern float3 __attribute__((overloadable)) powr(float3 v, float3 p) { 552 return pow(v, p); 553 } 554 extern float4 __attribute__((overloadable)) powr(float4 v, float4 p) { 555 return pow(v, p); 556 } 557 558 extern float __attribute__((overloadable)) remainder(float, float); 559 FN_FUNC_FN_FN(remainder) 560 561 extern float __attribute__((overloadable)) remquo(float, float, int *); 562 FN_FUNC_FN_FN_PIN(remquo) 563 564 extern float __attribute__((overloadable)) rint(float); 565 FN_FUNC_FN(rint) 566 567 extern float __attribute__((overloadable)) rootn(float v, int r) { 568 if (r == 0) { 569 return nan(0); 570 } 571 572 if (iszero(v)) { 573 if (r < 0) { 574 if (r & 1) { 575 return copysign(posinf(), v); 576 } else { 577 return posinf(); 578 } 579 } else { 580 if (r & 1) { 581 return copysign(0.f, v); 582 } else { 583 return 0.f; 584 } 585 } 586 } 587 588 if (!isinf(v) && !isnan(v) && (v < 0.f)) { 589 if (r & 1) { 590 return (-1.f * pow(-1.f * v, 1.f / r)); 591 } else { 592 return nan(0); 593 } 594 } 595 596 return pow(v, 1.f / r); 597 } 598 FN_FUNC_FN_IN(rootn); 599 600 extern float __attribute__((overloadable)) round(float); 601 FN_FUNC_FN(round) 602 603 604 extern float __attribute__((overloadable)) sqrt(float); 605 extern float __attribute__((overloadable)) rsqrt(float v) { 606 return 1.f / sqrt(v); 607 } 608 FN_FUNC_FN(rsqrt) 609 610 extern float __attribute__((overloadable)) sin(float); 611 FN_FUNC_FN(sin) 612 613 extern float __attribute__((overloadable)) sincos(float v, float *cosptr) { 614 *cosptr = cos(v); 615 return sin(v); 616 } 617 extern float2 __attribute__((overloadable)) sincos(float2 v, float2 *cosptr) { 618 *cosptr = cos(v); 619 return sin(v); 620 } 621 extern float3 __attribute__((overloadable)) sincos(float3 v, float3 *cosptr) { 622 *cosptr = cos(v); 623 return sin(v); 624 } 625 extern float4 __attribute__((overloadable)) sincos(float4 v, float4 *cosptr) { 626 *cosptr = cos(v); 627 return sin(v); 628 } 629 630 extern float __attribute__((overloadable)) sinh(float); 631 FN_FUNC_FN(sinh) 632 633 extern float __attribute__((overloadable)) sinpi(float v) { 634 return sin(v * M_PI); 635 } 636 FN_FUNC_FN(sinpi) 637 638 FN_FUNC_FN(sqrt) 639 640 extern float __attribute__((overloadable)) tan(float); 641 FN_FUNC_FN(tan) 642 643 extern float __attribute__((overloadable)) tanh(float); 644 FN_FUNC_FN(tanh) 645 646 extern float __attribute__((overloadable)) tanpi(float v) { 647 return tan(v * M_PI); 648 } 649 FN_FUNC_FN(tanpi) 650 651 652 extern float __attribute__((overloadable)) tgamma(float); 653 FN_FUNC_FN(tgamma) 654 655 extern float __attribute__((overloadable)) trunc(float); 656 FN_FUNC_FN(trunc) 657 658 // Int ops (partial), 6.11.3 659 660 #define XN_FUNC_YN(typeout, fnc, typein) \ 661 extern typeout __attribute__((overloadable)) fnc(typein); \ 662 extern typeout##2 __attribute__((overloadable)) fnc(typein##2 v) { \ 663 typeout##2 r; \ 664 r.x = fnc(v.x); \ 665 r.y = fnc(v.y); \ 666 return r; \ 667 } \ 668 extern typeout##3 __attribute__((overloadable)) fnc(typein##3 v) { \ 669 typeout##3 r; \ 670 r.x = fnc(v.x); \ 671 r.y = fnc(v.y); \ 672 r.z = fnc(v.z); \ 673 return r; \ 674 } \ 675 extern typeout##4 __attribute__((overloadable)) fnc(typein##4 v) { \ 676 typeout##4 r; \ 677 r.x = fnc(v.x); \ 678 r.y = fnc(v.y); \ 679 r.z = fnc(v.z); \ 680 r.w = fnc(v.w); \ 681 return r; \ 682 } 683 684 685 #define UIN_FUNC_IN(fnc) \ 686 XN_FUNC_YN(uchar, fnc, char) \ 687 XN_FUNC_YN(ushort, fnc, short) \ 688 XN_FUNC_YN(uint, fnc, int) 689 690 #define IN_FUNC_IN(fnc) \ 691 XN_FUNC_YN(uchar, fnc, uchar) \ 692 XN_FUNC_YN(char, fnc, char) \ 693 XN_FUNC_YN(ushort, fnc, ushort) \ 694 XN_FUNC_YN(short, fnc, short) \ 695 XN_FUNC_YN(uint, fnc, uint) \ 696 XN_FUNC_YN(int, fnc, int) 697 698 699 #define XN_FUNC_XN_XN_BODY(type, fnc, body) \ 700 extern type __attribute__((overloadable)) \ 701 fnc(type v1, type v2) { \ 702 return body; \ 703 } \ 704 extern type##2 __attribute__((overloadable)) \ 705 fnc(type##2 v1, type##2 v2) { \ 706 type##2 r; \ 707 r.x = fnc(v1.x, v2.x); \ 708 r.y = fnc(v1.y, v2.y); \ 709 return r; \ 710 } \ 711 extern type##3 __attribute__((overloadable)) \ 712 fnc(type##3 v1, type##3 v2) { \ 713 type##3 r; \ 714 r.x = fnc(v1.x, v2.x); \ 715 r.y = fnc(v1.y, v2.y); \ 716 r.z = fnc(v1.z, v2.z); \ 717 return r; \ 718 } \ 719 extern type##4 __attribute__((overloadable)) \ 720 fnc(type##4 v1, type##4 v2) { \ 721 type##4 r; \ 722 r.x = fnc(v1.x, v2.x); \ 723 r.y = fnc(v1.y, v2.y); \ 724 r.z = fnc(v1.z, v2.z); \ 725 r.w = fnc(v1.w, v2.w); \ 726 return r; \ 727 } 728 729 #define IN_FUNC_IN_IN_BODY(fnc, body) \ 730 XN_FUNC_XN_XN_BODY(uchar, fnc, body) \ 731 XN_FUNC_XN_XN_BODY(char, fnc, body) \ 732 XN_FUNC_XN_XN_BODY(ushort, fnc, body) \ 733 XN_FUNC_XN_XN_BODY(short, fnc, body) \ 734 XN_FUNC_XN_XN_BODY(uint, fnc, body) \ 735 XN_FUNC_XN_XN_BODY(int, fnc, body) \ 736 XN_FUNC_XN_XN_BODY(float, fnc, body) 737 738 UIN_FUNC_IN(abs) 739 IN_FUNC_IN(clz) 740 741 IN_FUNC_IN_IN_BODY(min, (v1 < v2 ? v1 : v2)) 742 FN_FUNC_FN_F(min) 743 744 IN_FUNC_IN_IN_BODY(max, (v1 > v2 ? v1 : v2)) 745 FN_FUNC_FN_F(max) 746 747 748 // 6.11.4 749 750 751 extern float __attribute__((overloadable)) degrees(float radians) { 752 return radians * (180.f / M_PI); 753 } 754 FN_FUNC_FN(degrees) 755 756 extern float __attribute__((overloadable)) mix(float start, float stop, float amount) { 757 return start + (stop - start) * amount; 758 } 759 extern float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float2 amount) { 760 return start + (stop - start) * amount; 761 } 762 extern float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float3 amount) { 763 return start + (stop - start) * amount; 764 } 765 extern float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float4 amount) { 766 return start + (stop - start) * amount; 767 } 768 extern float2 __attribute__((overloadable)) mix(float2 start, float2 stop, float amount) { 769 return start + (stop - start) * amount; 770 } 771 extern float3 __attribute__((overloadable)) mix(float3 start, float3 stop, float amount) { 772 return start + (stop - start) * amount; 773 } 774 extern float4 __attribute__((overloadable)) mix(float4 start, float4 stop, float amount) { 775 return start + (stop - start) * amount; 776 } 777 778 extern float __attribute__((overloadable)) radians(float degrees) { 779 return degrees * (M_PI / 180.f); 780 } 781 FN_FUNC_FN(radians) 782 783 extern float __attribute__((overloadable)) step(float edge, float v) { 784 return (v < edge) ? 0.f : 1.f; 785 } 786 extern float2 __attribute__((overloadable)) step(float2 edge, float2 v) { 787 float2 r; 788 r.x = (v.x < edge.x) ? 0.f : 1.f; 789 r.y = (v.y < edge.y) ? 0.f : 1.f; 790 return r; 791 } 792 extern float3 __attribute__((overloadable)) step(float3 edge, float3 v) { 793 float3 r; 794 r.x = (v.x < edge.x) ? 0.f : 1.f; 795 r.y = (v.y < edge.y) ? 0.f : 1.f; 796 r.z = (v.z < edge.z) ? 0.f : 1.f; 797 return r; 798 } 799 extern float4 __attribute__((overloadable)) step(float4 edge, float4 v) { 800 float4 r; 801 r.x = (v.x < edge.x) ? 0.f : 1.f; 802 r.y = (v.y < edge.y) ? 0.f : 1.f; 803 r.z = (v.z < edge.z) ? 0.f : 1.f; 804 r.w = (v.w < edge.w) ? 0.f : 1.f; 805 return r; 806 } 807 extern float2 __attribute__((overloadable)) step(float2 edge, float v) { 808 float2 r; 809 r.x = (v < edge.x) ? 0.f : 1.f; 810 r.y = (v < edge.y) ? 0.f : 1.f; 811 return r; 812 } 813 extern float3 __attribute__((overloadable)) step(float3 edge, float v) { 814 float3 r; 815 r.x = (v < edge.x) ? 0.f : 1.f; 816 r.y = (v < edge.y) ? 0.f : 1.f; 817 r.z = (v < edge.z) ? 0.f : 1.f; 818 return r; 819 } 820 extern float4 __attribute__((overloadable)) step(float4 edge, float v) { 821 float4 r; 822 r.x = (v < edge.x) ? 0.f : 1.f; 823 r.y = (v < edge.y) ? 0.f : 1.f; 824 r.z = (v < edge.z) ? 0.f : 1.f; 825 r.w = (v < edge.w) ? 0.f : 1.f; 826 return r; 827 } 828 829 extern float __attribute__((overloadable)) smoothstep(float, float, float); 830 extern float2 __attribute__((overloadable)) smoothstep(float2, float2, float2); 831 extern float3 __attribute__((overloadable)) smoothstep(float3, float3, float3); 832 extern float4 __attribute__((overloadable)) smoothstep(float4, float4, float4); 833 extern float2 __attribute__((overloadable)) smoothstep(float, float, float2); 834 extern float3 __attribute__((overloadable)) smoothstep(float, float, float3); 835 extern float4 __attribute__((overloadable)) smoothstep(float, float, float4); 836 837 extern float __attribute__((overloadable)) sign(float v) { 838 if (v > 0) return 1.f; 839 if (v < 0) return -1.f; 840 return v; 841 } 842 FN_FUNC_FN(sign) 843 844 845 // 6.11.5 846 extern float3 __attribute__((overloadable)) cross(float3 lhs, float3 rhs) { 847 float3 r; 848 r.x = lhs.y * rhs.z - lhs.z * rhs.y; 849 r.y = lhs.z * rhs.x - lhs.x * rhs.z; 850 r.z = lhs.x * rhs.y - lhs.y * rhs.x; 851 return r; 852 } 853 854 extern float4 __attribute__((overloadable)) cross(float4 lhs, float4 rhs) { 855 float4 r; 856 r.x = lhs.y * rhs.z - lhs.z * rhs.y; 857 r.y = lhs.z * rhs.x - lhs.x * rhs.z; 858 r.z = lhs.x * rhs.y - lhs.y * rhs.x; 859 r.w = 0.f; 860 return r; 861 } 862 863 extern float __attribute__((overloadable)) dot(float lhs, float rhs) { 864 return lhs * rhs; 865 } 866 extern float __attribute__((overloadable)) dot(float2 lhs, float2 rhs) { 867 return lhs.x*rhs.x + lhs.y*rhs.y; 868 } 869 extern float __attribute__((overloadable)) dot(float3 lhs, float3 rhs) { 870 return lhs.x*rhs.x + lhs.y*rhs.y + lhs.z*rhs.z; 871 } 872 extern float __attribute__((overloadable)) dot(float4 lhs, float4 rhs) { 873 return lhs.x*rhs.x + lhs.y*rhs.y + lhs.z*rhs.z + lhs.w*rhs.w; 874 } 875 876 extern float __attribute__((overloadable)) length(float v) { 877 return v; 878 } 879 extern float __attribute__((overloadable)) length(float2 v) { 880 return sqrt(v.x*v.x + v.y*v.y); 881 } 882 extern float __attribute__((overloadable)) length(float3 v) { 883 return sqrt(v.x*v.x + v.y*v.y + v.z*v.z); 884 } 885 extern float __attribute__((overloadable)) length(float4 v) { 886 return sqrt(v.x*v.x + v.y*v.y + v.z*v.z + v.w*v.w); 887 } 888 889 extern float __attribute__((overloadable)) distance(float lhs, float rhs) { 890 return length(lhs - rhs); 891 } 892 extern float __attribute__((overloadable)) distance(float2 lhs, float2 rhs) { 893 return length(lhs - rhs); 894 } 895 extern float __attribute__((overloadable)) distance(float3 lhs, float3 rhs) { 896 return length(lhs - rhs); 897 } 898 extern float __attribute__((overloadable)) distance(float4 lhs, float4 rhs) { 899 return length(lhs - rhs); 900 } 901 902 extern float __attribute__((overloadable)) normalize(float v) { 903 return 1.f; 904 } 905 extern float2 __attribute__((overloadable)) normalize(float2 v) { 906 return v / length(v); 907 } 908 extern float3 __attribute__((overloadable)) normalize(float3 v) { 909 return v / length(v); 910 } 911 extern float4 __attribute__((overloadable)) normalize(float4 v) { 912 return v / length(v); 913 } 914 915 #undef CVT_FUNC 916 #undef CVT_FUNC_2 917 #undef FN_FUNC_FN 918 #undef IN_FUNC_FN 919 #undef FN_FUNC_FN_FN 920 #undef FN_FUNC_FN_F 921 #undef FN_FUNC_FN_IN 922 #undef FN_FUNC_FN_I 923 #undef FN_FUNC_FN_PFN 924 #undef FN_FUNC_FN_PIN 925 #undef FN_FUNC_FN_FN_FN 926 #undef FN_FUNC_FN_FN_PIN 927 #undef XN_FUNC_YN 928 #undef UIN_FUNC_IN 929 #undef IN_FUNC_IN 930 #undef XN_FUNC_XN_XN_BODY 931 #undef IN_FUNC_IN_IN_BODY 932
