1 /////////////////////////////////////////////////////////////////////////////////// 2 /// OpenGL Mathematics (glm.g-truc.net) 3 /// 4 /// Copyright (c) 2005 - 2014 G-Truc Creation (www.g-truc.net) 5 /// Permission is hereby granted, free of charge, to any person obtaining a copy 6 /// of this software and associated documentation files (the "Software"), to deal 7 /// in the Software without restriction, including without limitation the rights 8 /// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 /// copies of the Software, and to permit persons to whom the Software is 10 /// furnished to do so, subject to the following conditions: 11 /// 12 /// The above copyright notice and this permission notice shall be included in 13 /// all copies or substantial portions of the Software. 14 /// 15 /// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 /// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 /// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 /// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 /// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 /// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 21 /// THE SOFTWARE. 22 /// 23 /// @ref gtx_simd_vec4 24 /// @file glm/gtx/simd_vec4.hpp 25 /// @date 2009-05-07 / 2011-06-07 26 /// @author Christophe Riccio 27 /// 28 /// @see core (dependence) 29 /// 30 /// @defgroup gtx_simd_vec4 GLM_GTX_simd_vec4 31 /// @ingroup gtx 32 /// 33 /// @brief SIMD implementation of vec4 type. 34 /// 35 /// <glm/gtx/simd_vec4.hpp> need to be included to use these functionalities. 36 /////////////////////////////////////////////////////////////////////////////////// 37 38 #ifndef GLM_GTX_simd_vec4 39 #define GLM_GTX_simd_vec4 40 41 // Dependency: 42 #include "../glm.hpp" 43 44 #if(GLM_ARCH != GLM_ARCH_PURE) 45 46 #if(GLM_ARCH & GLM_ARCH_SSE2) 47 # include "../detail/intrinsic_common.hpp" 48 # include "../detail/intrinsic_geometric.hpp" 49 # include "../detail/intrinsic_integer.hpp" 50 #else 51 # error "GLM: GLM_GTX_simd_vec4 requires compiler support of SSE2 through intrinsics" 52 #endif 53 54 #if(defined(GLM_MESSAGES) && !defined(GLM_EXT_INCLUDED)) 55 # pragma message("GLM: GLM_GTX_simd_vec4 extension included") 56 #endif 57 58 59 // Warning silencer for nameless struct/union. 60 #if (GLM_COMPILER & GLM_COMPILER_VC) 61 # pragma warning(push) 62 # pragma warning(disable:4201) // warning C4201: nonstandard extension used : nameless struct/union 63 #endif 64 65 namespace glm 66 { 67 enum comp 68 { 69 X = 0, 70 R = 0, 71 S = 0, 72 Y = 1, 73 G = 1, 74 T = 1, 75 Z = 2, 76 B = 2, 77 P = 2, 78 W = 3, 79 A = 3, 80 Q = 3 81 }; 82 83 }//namespace glm 84 85 namespace glm{ 86 namespace detail 87 { 88 /// 4-dimensional vector implemented using SIMD SEE intrinsics. 89 /// \ingroup gtx_simd_vec4 90 GLM_ALIGNED_STRUCT(16) fvec4SIMD 91 { 92 enum ctor{null}; 93 typedef __m128 value_type; 94 typedef std::size_t size_type; 95 static size_type value_size(); 96 97 typedef fvec4SIMD type; 98 typedef tvec4<bool, highp> bool_type; 99 100 #ifdef GLM_SIMD_ENABLE_XYZW_UNION 101 union 102 { 103 __m128 Data; 104 struct {float x, y, z, w;}; 105 }; 106 #else 107 __m128 Data; 108 #endif 109 110 ////////////////////////////////////// 111 // Implicit basic constructors 112 113 fvec4SIMD(); 114 fvec4SIMD(__m128 const & Data); 115 fvec4SIMD(fvec4SIMD const & v); 116 117 ////////////////////////////////////// 118 // Explicit basic constructors 119 120 explicit fvec4SIMD( 121 ctor); 122 explicit fvec4SIMD( 123 float const & s); 124 explicit fvec4SIMD( 125 float const & x, 126 float const & y, 127 float const & z, 128 float const & w); 129 explicit fvec4SIMD( 130 vec4 const & v); 131 132 //////////////////////////////////////// 133 //// Conversion vector constructors 134 135 fvec4SIMD(vec2 const & v, float const & s1, float const & s2); 136 fvec4SIMD(float const & s1, vec2 const & v, float const & s2); 137 fvec4SIMD(float const & s1, float const & s2, vec2 const & v); 138 fvec4SIMD(vec3 const & v, float const & s); 139 fvec4SIMD(float const & s, vec3 const & v); 140 fvec4SIMD(vec2 const & v1, vec2 const & v2); 141 //fvec4SIMD(ivec4SIMD const & v); 142 143 ////////////////////////////////////// 144 // Unary arithmetic operators 145 146 fvec4SIMD& operator= (fvec4SIMD const & v); 147 fvec4SIMD& operator+=(fvec4SIMD const & v); 148 fvec4SIMD& operator-=(fvec4SIMD const & v); 149 fvec4SIMD& operator*=(fvec4SIMD const & v); 150 fvec4SIMD& operator/=(fvec4SIMD const & v); 151 152 fvec4SIMD& operator+=(float const & s); 153 fvec4SIMD& operator-=(float const & s); 154 fvec4SIMD& operator*=(float const & s); 155 fvec4SIMD& operator/=(float const & s); 156 157 fvec4SIMD& operator++(); 158 fvec4SIMD& operator--(); 159 160 ////////////////////////////////////// 161 // Swizzle operators 162 163 template <comp X, comp Y, comp Z, comp W> 164 fvec4SIMD& swizzle(); 165 template <comp X, comp Y, comp Z, comp W> 166 fvec4SIMD swizzle() const; 167 template <comp X, comp Y, comp Z> 168 fvec4SIMD swizzle() const; 169 template <comp X, comp Y> 170 fvec4SIMD swizzle() const; 171 template <comp X> 172 fvec4SIMD swizzle() const; 173 }; 174 }//namespace detail 175 176 typedef glm::detail::fvec4SIMD simdVec4; 177 178 /// @addtogroup gtx_simd_vec4 179 /// @{ 180 181 //! Convert a simdVec4 to a vec4. 182 /// @see gtx_simd_vec4 183 vec4 vec4_cast( 184 detail::fvec4SIMD const & x); 185 186 //! Returns x if x >= 0; otherwise, it returns -x. 187 /// @see gtx_simd_vec4 188 detail::fvec4SIMD abs(detail::fvec4SIMD const & x); 189 190 //! Returns 1.0 if x > 0, 0.0 if x = 0, or -1.0 if x < 0. 191 /// @see gtx_simd_vec4 192 detail::fvec4SIMD sign(detail::fvec4SIMD const & x); 193 194 //! Returns a value equal to the nearest integer that is less then or equal to x. 195 /// @see gtx_simd_vec4 196 detail::fvec4SIMD floor(detail::fvec4SIMD const & x); 197 198 //! Returns a value equal to the nearest integer to x 199 //! whose absolute value is not larger than the absolute value of x. 200 /// @see gtx_simd_vec4 201 detail::fvec4SIMD trunc(detail::fvec4SIMD const & x); 202 203 //! Returns a value equal to the nearest integer to x. 204 //! The fraction 0.5 will round in a direction chosen by the 205 //! implementation, presumably the direction that is fastest. 206 //! This includes the possibility that round(x) returns the 207 //! same value as roundEven(x) for all values of x. 208 /// 209 /// @see gtx_simd_vec4 210 detail::fvec4SIMD round(detail::fvec4SIMD const & x); 211 212 //! Returns a value equal to the nearest integer to x. 213 //! A fractional part of 0.5 will round toward the nearest even 214 //! integer. (Both 3.5 and 4.5 for x will return 4.0.) 215 /// 216 /// @see gtx_simd_vec4 217 //detail::fvec4SIMD roundEven(detail::fvec4SIMD const & x); 218 219 //! Returns a value equal to the nearest integer 220 //! that is greater than or equal to x. 221 /// @see gtx_simd_vec4 222 detail::fvec4SIMD ceil(detail::fvec4SIMD const & x); 223 224 //! Return x - floor(x). 225 /// 226 /// @see gtx_simd_vec4 227 detail::fvec4SIMD fract(detail::fvec4SIMD const & x); 228 229 //! Modulus. Returns x - y * floor(x / y) 230 //! for each component in x using the floating point value y. 231 /// 232 /// @see gtx_simd_vec4 233 detail::fvec4SIMD mod( 234 detail::fvec4SIMD const & x, 235 detail::fvec4SIMD const & y); 236 237 //! Modulus. Returns x - y * floor(x / y) 238 //! for each component in x using the floating point value y. 239 /// 240 /// @see gtx_simd_vec4 241 detail::fvec4SIMD mod( 242 detail::fvec4SIMD const & x, 243 float const & y); 244 245 //! Returns the fractional part of x and sets i to the integer 246 //! part (as a whole number floating point value). Both the 247 //! return value and the output parameter will have the same 248 //! sign as x. 249 //! (From GLM_GTX_simd_vec4 extension, common function) 250 //detail::fvec4SIMD modf( 251 // detail::fvec4SIMD const & x, 252 // detail::fvec4SIMD & i); 253 254 //! Returns y if y < x; otherwise, it returns x. 255 /// 256 /// @see gtx_simd_vec4 257 detail::fvec4SIMD min( 258 detail::fvec4SIMD const & x, 259 detail::fvec4SIMD const & y); 260 261 detail::fvec4SIMD min( 262 detail::fvec4SIMD const & x, 263 float const & y); 264 265 //! Returns y if x < y; otherwise, it returns x. 266 /// 267 /// @see gtx_simd_vec4 268 detail::fvec4SIMD max( 269 detail::fvec4SIMD const & x, 270 detail::fvec4SIMD const & y); 271 272 detail::fvec4SIMD max( 273 detail::fvec4SIMD const & x, 274 float const & y); 275 276 //! Returns min(max(x, minVal), maxVal) for each component in x 277 //! using the floating-point values minVal and maxVal. 278 /// 279 /// @see gtx_simd_vec4 280 detail::fvec4SIMD clamp( 281 detail::fvec4SIMD const & x, 282 detail::fvec4SIMD const & minVal, 283 detail::fvec4SIMD const & maxVal); 284 285 detail::fvec4SIMD clamp( 286 detail::fvec4SIMD const & x, 287 float const & minVal, 288 float const & maxVal); 289 290 //! \return If genTypeU is a floating scalar or vector: 291 //! Returns x * (1.0 - a) + y * a, i.e., the linear blend of 292 //! x and y using the floating-point value a. 293 //! The value for a is not restricted to the range [0, 1]. 294 //! 295 //! \return If genTypeU is a boolean scalar or vector: 296 //! Selects which vector each returned component comes 297 //! from. For a component of a that is false, the 298 //! corresponding component of x is returned. For a 299 //! component of a that is true, the corresponding 300 //! component of y is returned. Components of x and y that 301 //! are not selected are allowed to be invalid floating point 302 //! values and will have no effect on the results. Thus, this 303 //! provides different functionality than 304 //! genType mix(genType x, genType y, genType(a)) 305 //! where a is a Boolean vector. 306 //! 307 //! From GLSL 1.30.08 specification, section 8.3 308 //! 309 //! \param[in] x Floating point scalar or vector. 310 //! \param[in] y Floating point scalar or vector. 311 //! \param[in] a Floating point or boolean scalar or vector. 312 //! 313 /// \todo Test when 'a' is a boolean. 314 /// 315 /// @see gtx_simd_vec4 316 detail::fvec4SIMD mix( 317 detail::fvec4SIMD const & x, 318 detail::fvec4SIMD const & y, 319 detail::fvec4SIMD const & a); 320 321 //! Returns 0.0 if x < edge, otherwise it returns 1.0. 322 /// 323 /// @see gtx_simd_vec4 324 detail::fvec4SIMD step( 325 detail::fvec4SIMD const & edge, 326 detail::fvec4SIMD const & x); 327 328 detail::fvec4SIMD step( 329 float const & edge, 330 detail::fvec4SIMD const & x); 331 332 //! Returns 0.0 if x <= edge0 and 1.0 if x >= edge1 and 333 //! performs smooth Hermite interpolation between 0 and 1 334 //! when edge0 < x < edge1. This is useful in cases where 335 //! you would want a threshold function with a smooth 336 //! transition. This is equivalent to: 337 //! genType t; 338 //! t = clamp ((x - edge0) / (edge1 - edge0), 0, 1); 339 //! return t * t * (3 - 2 * t); 340 //! Results are undefined if edge0 >= edge1. 341 /// 342 /// @see gtx_simd_vec4 343 detail::fvec4SIMD smoothstep( 344 detail::fvec4SIMD const & edge0, 345 detail::fvec4SIMD const & edge1, 346 detail::fvec4SIMD const & x); 347 348 detail::fvec4SIMD smoothstep( 349 float const & edge0, 350 float const & edge1, 351 detail::fvec4SIMD const & x); 352 353 //! Returns true if x holds a NaN (not a number) 354 //! representation in the underlying implementation's set of 355 //! floating point representations. Returns false otherwise, 356 //! including for implementations with no NaN 357 //! representations. 358 /// 359 /// @see gtx_simd_vec4 360 //bvec4 isnan(detail::fvec4SIMD const & x); 361 362 //! Returns true if x holds a positive infinity or negative 363 //! infinity representation in the underlying implementation's 364 //! set of floating point representations. Returns false 365 //! otherwise, including for implementations with no infinity 366 //! representations. 367 /// 368 /// @see gtx_simd_vec4 369 //bvec4 isinf(detail::fvec4SIMD const & x); 370 371 //! Returns a signed or unsigned integer value representing 372 //! the encoding of a floating-point value. The floatingpoint 373 //! value's bit-level representation is preserved. 374 /// 375 /// @see gtx_simd_vec4 376 //detail::ivec4SIMD floatBitsToInt(detail::fvec4SIMD const & value); 377 378 //! Returns a floating-point value corresponding to a signed 379 //! or unsigned integer encoding of a floating-point value. 380 //! If an inf or NaN is passed in, it will not signal, and the 381 //! resulting floating point value is unspecified. Otherwise, 382 //! the bit-level representation is preserved. 383 /// 384 /// @see gtx_simd_vec4 385 //detail::fvec4SIMD intBitsToFloat(detail::ivec4SIMD const & value); 386 387 //! Computes and returns a * b + c. 388 /// 389 /// @see gtx_simd_vec4 390 detail::fvec4SIMD fma( 391 detail::fvec4SIMD const & a, 392 detail::fvec4SIMD const & b, 393 detail::fvec4SIMD const & c); 394 395 //! Splits x into a floating-point significand in the range 396 //! [0.5, 1.0) and an integral exponent of two, such that: 397 //! x = significand * exp(2, exponent) 398 //! The significand is returned by the function and the 399 //! exponent is returned in the parameter exp. For a 400 //! floating-point value of zero, the significant and exponent 401 //! are both zero. For a floating-point value that is an 402 //! infinity or is not a number, the results are undefined. 403 /// 404 /// @see gtx_simd_vec4 405 //detail::fvec4SIMD frexp(detail::fvec4SIMD const & x, detail::ivec4SIMD & exp); 406 407 //! Builds a floating-point number from x and the 408 //! corresponding integral exponent of two in exp, returning: 409 //! significand * exp(2, exponent) 410 //! If this product is too large to be represented in the 411 //! floating-point type, the result is undefined. 412 /// 413 /// @see gtx_simd_vec4 414 //detail::fvec4SIMD ldexp(detail::fvec4SIMD const & x, detail::ivec4SIMD const & exp); 415 416 //! Returns the length of x, i.e., sqrt(x * x). 417 /// 418 /// @see gtx_simd_vec4 419 float length( 420 detail::fvec4SIMD const & x); 421 422 //! Returns the length of x, i.e., sqrt(x * x). 423 //! Less accurate but much faster than simdLength. 424 /// 425 /// @see gtx_simd_vec4 426 float fastLength( 427 detail::fvec4SIMD const & x); 428 429 //! Returns the length of x, i.e., sqrt(x * x). 430 //! Slightly more accurate but much slower than simdLength. 431 /// 432 /// @see gtx_simd_vec4 433 float niceLength( 434 detail::fvec4SIMD const & x); 435 436 //! Returns the length of x, i.e., sqrt(x * x). 437 /// 438 /// @see gtx_simd_vec4 439 detail::fvec4SIMD length4( 440 detail::fvec4SIMD const & x); 441 442 //! Returns the length of x, i.e., sqrt(x * x). 443 //! Less accurate but much faster than simdLength4. 444 /// 445 /// @see gtx_simd_vec4 446 detail::fvec4SIMD fastLength4( 447 detail::fvec4SIMD const & x); 448 449 //! Returns the length of x, i.e., sqrt(x * x). 450 //! Slightly more accurate but much slower than simdLength4. 451 /// 452 /// @see gtx_simd_vec4 453 detail::fvec4SIMD niceLength4( 454 detail::fvec4SIMD const & x); 455 456 //! Returns the distance betwwen p0 and p1, i.e., length(p0 - p1). 457 /// 458 /// @see gtx_simd_vec4 459 float distance( 460 detail::fvec4SIMD const & p0, 461 detail::fvec4SIMD const & p1); 462 463 //! Returns the distance betwwen p0 and p1, i.e., length(p0 - p1). 464 /// 465 /// @see gtx_simd_vec4 466 detail::fvec4SIMD distance4( 467 detail::fvec4SIMD const & p0, 468 detail::fvec4SIMD const & p1); 469 470 //! Returns the dot product of x and y, i.e., result = x * y. 471 /// 472 /// @see gtx_simd_vec4 473 float simdDot( 474 detail::fvec4SIMD const & x, 475 detail::fvec4SIMD const & y); 476 477 //! Returns the dot product of x and y, i.e., result = x * y. 478 /// 479 /// @see gtx_simd_vec4 480 detail::fvec4SIMD dot4( 481 detail::fvec4SIMD const & x, 482 detail::fvec4SIMD const & y); 483 484 //! Returns the cross product of x and y. 485 /// 486 /// @see gtx_simd_vec4 487 detail::fvec4SIMD cross( 488 detail::fvec4SIMD const & x, 489 detail::fvec4SIMD const & y); 490 491 //! Returns a vector in the same direction as x but with length of 1. 492 /// 493 /// @see gtx_simd_vec4 494 detail::fvec4SIMD normalize( 495 detail::fvec4SIMD const & x); 496 497 //! Returns a vector in the same direction as x but with length of 1. 498 //! Less accurate but much faster than simdNormalize. 499 /// 500 /// @see gtx_simd_vec4 501 detail::fvec4SIMD fastNormalize( 502 detail::fvec4SIMD const & x); 503 504 //! If dot(Nref, I) < 0.0, return N, otherwise, return -N. 505 /// 506 /// @see gtx_simd_vec4 507 detail::fvec4SIMD simdFaceforward( 508 detail::fvec4SIMD const & N, 509 detail::fvec4SIMD const & I, 510 detail::fvec4SIMD const & Nref); 511 512 //! For the incident vector I and surface orientation N, 513 //! returns the reflection direction : result = I - 2.0 * dot(N, I) * N. 514 /// 515 /// @see gtx_simd_vec4 516 detail::fvec4SIMD reflect( 517 detail::fvec4SIMD const & I, 518 detail::fvec4SIMD const & N); 519 520 //! For the incident vector I and surface normal N, 521 //! and the ratio of indices of refraction eta, 522 //! return the refraction vector. 523 /// 524 /// @see gtx_simd_vec4 525 detail::fvec4SIMD refract( 526 detail::fvec4SIMD const & I, 527 detail::fvec4SIMD const & N, 528 float const & eta); 529 530 //! Returns the positive square root of x. 531 /// 532 /// @see gtx_simd_vec4 533 detail::fvec4SIMD sqrt( 534 detail::fvec4SIMD const & x); 535 536 //! Returns the positive square root of x with the nicest quality but very slow. 537 //! Slightly more accurate but much slower than simdSqrt. 538 /// 539 /// @see gtx_simd_vec4 540 detail::fvec4SIMD niceSqrt( 541 detail::fvec4SIMD const & x); 542 543 //! Returns the positive square root of x 544 //! Less accurate but much faster than sqrt. 545 /// 546 /// @see gtx_simd_vec4 547 detail::fvec4SIMD fastSqrt( 548 detail::fvec4SIMD const & x); 549 550 //! Returns the reciprocal of the positive square root of x. 551 /// 552 /// @see gtx_simd_vec4 553 detail::fvec4SIMD inversesqrt( 554 detail::fvec4SIMD const & x); 555 556 //! Returns the reciprocal of the positive square root of x. 557 //! Faster than inversesqrt but less accurate. 558 /// 559 /// @see gtx_simd_vec4 560 detail::fvec4SIMD fastInversesqrt( 561 detail::fvec4SIMD const & x); 562 563 /// @} 564 }//namespace glm 565 566 #include "simd_vec4.inl" 567 568 #if (GLM_COMPILER & GLM_COMPILER_VC) 569 # pragma warning(pop) 570 #endif 571 572 #endif//(GLM_ARCH != GLM_ARCH_PURE) 573 574 #endif//GLM_GTX_simd_vec4 575
