2 Copyright (C) 2006, 2010 Sony Computer Entertainment Inc.
\r
5 Redistribution and use in source and binary forms,
\r
6 with or without modification, are permitted provided that the
\r
7 following conditions are met:
\r
8 * Redistributions of source code must retain the above copyright
\r
9 notice, this list of conditions and the following disclaimer.
\r
10 * Redistributions in binary form must reproduce the above copyright
\r
11 notice, this list of conditions and the following disclaimer in the
\r
12 documentation and/or other materials provided with the distribution.
\r
13 * Neither the name of the Sony Computer Entertainment Inc nor the names
\r
14 of its contributors may be used to endorse or promote products derived
\r
15 from this software without specific prior written permission.
\r
17 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
\r
18 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
\r
19 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
\r
20 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
\r
21 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
\r
22 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
\r
23 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
\r
24 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
\r
25 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
\r
26 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
\r
27 POSSIBILITY OF SUCH DAMAGE.
\r
31 #ifndef _VECTORMATH_AOS_CPP_SSE_H
\r
32 #define _VECTORMATH_AOS_CPP_SSE_H
\r
35 #include <xmmintrin.h>
\r
36 #include <emmintrin.h>
\r
39 #define USE_SSE2_LDDQU
\r
40 #ifdef USE_SSE2_LDDQU
\r
41 #include <intrin.h> //used for _mm_lddqu_si128
\r
42 #endif //USE_SSE2_LDDQU
\r
45 typedef __m128 vec_float4;
\r
46 typedef __m128 vec_uint4;
\r
47 typedef __m128 vec_int4;
\r
48 typedef __m128i vec_uchar16;
\r
49 typedef __m128i vec_ushort8;
\r
51 #define vec_splat(x, e) _mm_shuffle_ps(x, x, _MM_SHUFFLE(e,e,e,e))
\r
53 #define _mm_ror_ps(vec,i) \
\r
54 (((i)%4) ? (_mm_shuffle_ps(vec,vec, _MM_SHUFFLE((unsigned char)(i+3)%4,(unsigned char)(i+2)%4,(unsigned char)(i+1)%4,(unsigned char)(i+0)%4))) : (vec))
\r
55 #define _mm_rol_ps(vec,i) \
\r
56 (((i)%4) ? (_mm_shuffle_ps(vec,vec, _MM_SHUFFLE((unsigned char)(7-i)%4,(unsigned char)(6-i)%4,(unsigned char)(5-i)%4,(unsigned char)(4-i)%4))) : (vec))
\r
58 #define vec_sld(vec,vec2,x) _mm_ror_ps(vec, ((x)/4))
\r
60 #define _mm_abs_ps(vec) _mm_andnot_ps(_MASKSIGN_,vec)
\r
61 #define _mm_neg_ps(vec) _mm_xor_ps(_MASKSIGN_,vec)
\r
63 #define vec_madd(a, b, c) _mm_add_ps(c, _mm_mul_ps(a, b) )
\r
71 unsigned short s[8];
\r
73 SSEFloat(__m128 v) : m128(v) {}
\r
74 SSEFloat(__m128i v) : vi(v) {}
\r
75 SSEFloat() {}//uninitialized
\r
78 static __forceinline __m128 vec_sel(__m128 a, __m128 b, __m128 mask)
\r
80 return _mm_or_ps(_mm_and_ps(mask, b), _mm_andnot_ps(mask, a));
\r
82 static __forceinline __m128 vec_sel(__m128 a, __m128 b, const unsigned int *_mask)
\r
84 return vec_sel(a, b, _mm_load_ps((float *)_mask));
\r
86 static __forceinline __m128 vec_sel(__m128 a, __m128 b, unsigned int _mask)
\r
88 return vec_sel(a, b, _mm_set1_ps(*(float *)&_mask));
\r
91 static __forceinline __m128 toM128(unsigned int x)
\r
93 return _mm_set1_ps( *(float *)&x );
\r
96 static __forceinline __m128 fabsf4(__m128 x)
\r
98 return _mm_and_ps( x, toM128( 0x7fffffff ) );
\r
111 static __forceinline __m128 vec_cts(__m128 x, int a)
\r
113 assert(a == 0); // Only 2^0 supported
\r
116 sse64.m64.m01 = _mm_cvttps_pi32(x);
\r
117 sse64.m64.m23 = _mm_cvttps_pi32(_mm_ror_ps(x,2));
\r
122 static __forceinline __m128 vec_ctf(__m128 x, int a)
\r
124 assert(a == 0); // Only 2^0 supported
\r
128 __m128 result =_mm_movelh_ps(
\r
129 _mm_cvt_pi2ps(_mm_setzero_ps(), sse64.m64.m01),
\r
130 _mm_cvt_pi2ps(_mm_setzero_ps(), sse64.m64.m23));
\r
135 static __forceinline __m128 vec_cts(__m128 x, int a)
\r
137 assert(a == 0); // Only 2^0 supported
\r
139 __m128i result = _mm_cvtps_epi32(x);
\r
140 return (__m128 &)result;
\r
143 static __forceinline __m128 vec_ctf(__m128 x, int a)
\r
145 assert(a == 0); // Only 2^0 supported
\r
147 return _mm_cvtepi32_ps((__m128i &)x);
\r
150 #define vec_nmsub(a,b,c) _mm_sub_ps( c, _mm_mul_ps( a, b ) )
\r
151 #define vec_sub(a,b) _mm_sub_ps( a, b )
\r
152 #define vec_add(a,b) _mm_add_ps( a, b )
\r
153 #define vec_mul(a,b) _mm_mul_ps( a, b )
\r
154 #define vec_xor(a,b) _mm_xor_ps( a, b )
\r
155 #define vec_and(a,b) _mm_and_ps( a, b )
\r
156 #define vec_cmpeq(a,b) _mm_cmpeq_ps( a, b )
\r
157 #define vec_cmpgt(a,b) _mm_cmpgt_ps( a, b )
\r
159 #define vec_mergeh(a,b) _mm_unpacklo_ps( a, b )
\r
160 #define vec_mergel(a,b) _mm_unpackhi_ps( a, b )
\r
162 #define vec_andc(a,b) _mm_andnot_ps( b, a )
\r
164 #define sqrtf4(x) _mm_sqrt_ps( x )
\r
165 #define rsqrtf4(x) _mm_rsqrt_ps( x )
\r
166 #define recipf4(x) _mm_rcp_ps( x )
\r
167 #define negatef4(x) _mm_sub_ps( _mm_setzero_ps(), x )
\r
169 static __forceinline __m128 newtonrapson_rsqrt4( const __m128 v )
\r
171 #define _half4 _mm_setr_ps(.5f,.5f,.5f,.5f)
\r
172 #define _three _mm_setr_ps(3.f,3.f,3.f,3.f)
\r
173 const __m128 approx = _mm_rsqrt_ps( v );
\r
174 const __m128 muls = _mm_mul_ps(_mm_mul_ps(v, approx), approx);
\r
175 return _mm_mul_ps(_mm_mul_ps(_half4, approx), _mm_sub_ps(_three, muls) );
\r
178 static __forceinline __m128 acosf4(__m128 x)
\r
180 __m128 xabs = fabsf4(x);
\r
181 __m128 select = _mm_cmplt_ps( x, _mm_setzero_ps() );
\r
182 __m128 t1 = sqrtf4(vec_sub(_mm_set1_ps(1.0f), xabs));
\r
184 /* Instruction counts can be reduced if the polynomial was
\r
185 * computed entirely from nested (dependent) fma's. However,
\r
186 * to reduce the number of pipeline stalls, the polygon is evaluated
\r
187 * in two halves (hi amd lo).
\r
189 __m128 xabs2 = _mm_mul_ps(xabs, xabs);
\r
190 __m128 xabs4 = _mm_mul_ps(xabs2, xabs2);
\r
191 __m128 hi = vec_madd(vec_madd(vec_madd(_mm_set1_ps(-0.0012624911f),
\r
192 xabs, _mm_set1_ps(0.0066700901f)),
\r
193 xabs, _mm_set1_ps(-0.0170881256f)),
\r
194 xabs, _mm_set1_ps( 0.0308918810f));
\r
195 __m128 lo = vec_madd(vec_madd(vec_madd(_mm_set1_ps(-0.0501743046f),
\r
196 xabs, _mm_set1_ps(0.0889789874f)),
\r
197 xabs, _mm_set1_ps(-0.2145988016f)),
\r
198 xabs, _mm_set1_ps( 1.5707963050f));
\r
200 __m128 result = vec_madd(hi, xabs4, lo);
\r
202 // Adjust the result if x is negactive.
\r
204 vec_mul(t1, result), // Positive
\r
205 vec_nmsub(t1, result, _mm_set1_ps(3.1415926535898f)), // Negative
\r
209 static __forceinline __m128 sinf4(vec_float4 x)
\r
213 // Common constants used to evaluate sinf4/cosf4/tanf4
\r
215 #define _SINCOS_CC0 -0.0013602249f
\r
216 #define _SINCOS_CC1 0.0416566950f
\r
217 #define _SINCOS_CC2 -0.4999990225f
\r
218 #define _SINCOS_SC0 -0.0001950727f
\r
219 #define _SINCOS_SC1 0.0083320758f
\r
220 #define _SINCOS_SC2 -0.1666665247f
\r
222 #define _SINCOS_KC1 1.57079625129f
\r
223 #define _SINCOS_KC2 7.54978995489e-8f
\r
225 vec_float4 xl,xl2,xl3,res;
\r
227 // Range reduction using : xl = angle * TwoOverPi;
\r
229 xl = vec_mul(x, _mm_set1_ps(0.63661977236f));
\r
231 // Find the quadrant the angle falls in
\r
232 // using: q = (int) (ceil(abs(xl))*sign(xl))
\r
234 vec_int4 q = vec_cts(xl,0);
\r
236 // Compute an offset based on the quadrant that the angle falls in
\r
238 vec_int4 offset = _mm_and_ps(q,toM128(0x3));
\r
240 // Remainder in range [-pi/4..pi/4]
\r
242 vec_float4 qf = vec_ctf(q,0);
\r
243 xl = vec_nmsub(qf,_mm_set1_ps(_SINCOS_KC2),vec_nmsub(qf,_mm_set1_ps(_SINCOS_KC1),x));
\r
245 // Compute x^2 and x^3
\r
247 xl2 = vec_mul(xl,xl);
\r
248 xl3 = vec_mul(xl2,xl);
\r
250 // Compute both the sin and cos of the angles
\r
251 // using a polynomial expression:
\r
252 // cx = 1.0f + xl2 * ((C0 * xl2 + C1) * xl2 + C2), and
\r
253 // sx = xl + xl3 * ((S0 * xl2 + S1) * xl2 + S2)
\r
259 vec_madd(_mm_set1_ps(_SINCOS_CC0),xl2,_mm_set1_ps(_SINCOS_CC1)),xl2,_mm_set1_ps(_SINCOS_CC2)),xl2,_mm_set1_ps(1.0f));
\r
263 vec_madd(_mm_set1_ps(_SINCOS_SC0),xl2,_mm_set1_ps(_SINCOS_SC1)),xl2,_mm_set1_ps(_SINCOS_SC2)),xl3,xl);
\r
265 // Use the cosine when the offset is odd and the sin
\r
266 // when the offset is even
\r
268 res = vec_sel(cx,sx,vec_cmpeq(vec_and(offset,
\r
270 _mm_setzero_ps()));
\r
272 // Flip the sign of the result when (offset mod 4) = 1 or 2
\r
275 vec_xor(toM128(0x80000000U), res), // Negative
\r
277 vec_cmpeq(vec_and(offset,toM128(0x2)),_mm_setzero_ps()));
\r
280 static __forceinline void sincosf4(vec_float4 x, vec_float4* s, vec_float4* c)
\r
282 vec_float4 xl,xl2,xl3;
\r
283 vec_int4 offsetSin, offsetCos;
\r
285 // Range reduction using : xl = angle * TwoOverPi;
\r
287 xl = vec_mul(x, _mm_set1_ps(0.63661977236f));
\r
289 // Find the quadrant the angle falls in
\r
290 // using: q = (int) (ceil(abs(xl))*sign(xl))
\r
292 //vec_int4 q = vec_cts(vec_add(xl,vec_sel(_mm_set1_ps(0.5f),xl,(0x80000000))),0);
\r
293 vec_int4 q = vec_cts(xl,0);
\r
295 // Compute the offset based on the quadrant that the angle falls in.
\r
296 // Add 1 to the offset for the cosine.
\r
298 offsetSin = vec_and(q,toM128((int)0x3));
\r
299 __m128i temp = _mm_add_epi32(_mm_set1_epi32(1),(__m128i &)offsetSin);
\r
300 offsetCos = (__m128 &)temp;
\r
302 // Remainder in range [-pi/4..pi/4]
\r
304 vec_float4 qf = vec_ctf(q,0);
\r
305 xl = vec_nmsub(qf,_mm_set1_ps(_SINCOS_KC2),vec_nmsub(qf,_mm_set1_ps(_SINCOS_KC1),x));
\r
307 // Compute x^2 and x^3
\r
309 xl2 = vec_mul(xl,xl);
\r
310 xl3 = vec_mul(xl2,xl);
\r
312 // Compute both the sin and cos of the angles
\r
313 // using a polynomial expression:
\r
314 // cx = 1.0f + xl2 * ((C0 * xl2 + C1) * xl2 + C2), and
\r
315 // sx = xl + xl3 * ((S0 * xl2 + S1) * xl2 + S2)
\r
320 vec_madd(_mm_set1_ps(_SINCOS_CC0),xl2,_mm_set1_ps(_SINCOS_CC1)),xl2,_mm_set1_ps(_SINCOS_CC2)),xl2,_mm_set1_ps(1.0f));
\r
324 vec_madd(_mm_set1_ps(_SINCOS_SC0),xl2,_mm_set1_ps(_SINCOS_SC1)),xl2,_mm_set1_ps(_SINCOS_SC2)),xl3,xl);
\r
326 // Use the cosine when the offset is odd and the sin
\r
327 // when the offset is even
\r
329 vec_uint4 sinMask = (vec_uint4)vec_cmpeq(vec_and(offsetSin,toM128(0x1)),_mm_setzero_ps());
\r
330 vec_uint4 cosMask = (vec_uint4)vec_cmpeq(vec_and(offsetCos,toM128(0x1)),_mm_setzero_ps());
\r
331 *s = vec_sel(cx,sx,sinMask);
\r
332 *c = vec_sel(cx,sx,cosMask);
\r
334 // Flip the sign of the result when (offset mod 4) = 1 or 2
\r
336 sinMask = vec_cmpeq(vec_and(offsetSin,toM128(0x2)),_mm_setzero_ps());
\r
337 cosMask = vec_cmpeq(vec_and(offsetCos,toM128(0x2)),_mm_setzero_ps());
\r
339 *s = vec_sel((vec_float4)vec_xor(toM128(0x80000000),(vec_uint4)*s),*s,sinMask);
\r
340 *c = vec_sel((vec_float4)vec_xor(toM128(0x80000000),(vec_uint4)*c),*c,cosMask);
\r
343 #include "vecidx_aos.h"
\r
344 #include "floatInVec.h"
\r
345 #include "boolInVec.h"
\r
347 #ifdef _VECTORMATH_DEBUG
\r
350 namespace Vectormath {
\r
354 //-----------------------------------------------------------------------------
\r
355 // Forward Declarations
\r
366 // A 3-D vector in array-of-structures format
\r
372 __forceinline void set128(vec_float4 vec);
\r
374 __forceinline vec_float4& get128Ref();
\r
377 // Default constructor; does no initialization
\r
379 __forceinline Vector3( ) { };
\r
381 // Default copy constructor
\r
383 __forceinline Vector3(const Vector3& vec);
\r
385 // Construct a 3-D vector from x, y, and z elements
\r
387 __forceinline Vector3( float x, float y, float z );
\r
389 // Construct a 3-D vector from x, y, and z elements (scalar data contained in vector data type)
\r
391 __forceinline Vector3( const floatInVec &x, const floatInVec &y, const floatInVec &z );
\r
393 // Copy elements from a 3-D point into a 3-D vector
\r
395 explicit __forceinline Vector3( const Point3 &pnt );
\r
397 // Set all elements of a 3-D vector to the same scalar value
\r
399 explicit __forceinline Vector3( float scalar );
\r
401 // Set all elements of a 3-D vector to the same scalar value (scalar data contained in vector data type)
\r
403 explicit __forceinline Vector3( const floatInVec &scalar );
\r
405 // Set vector float data in a 3-D vector
\r
407 explicit __forceinline Vector3( __m128 vf4 );
\r
409 // Get vector float data from a 3-D vector
\r
411 __forceinline __m128 get128( ) const;
\r
413 // Assign one 3-D vector to another
\r
415 __forceinline Vector3 & operator =( const Vector3 &vec );
\r
417 // Set the x element of a 3-D vector
\r
419 __forceinline Vector3 & setX( float x );
\r
421 // Set the y element of a 3-D vector
\r
423 __forceinline Vector3 & setY( float y );
\r
425 // Set the z element of a 3-D vector
\r
427 __forceinline Vector3 & setZ( float z );
\r
429 // Set the x element of a 3-D vector (scalar data contained in vector data type)
\r
431 __forceinline Vector3 & setX( const floatInVec &x );
\r
433 // Set the y element of a 3-D vector (scalar data contained in vector data type)
\r
435 __forceinline Vector3 & setY( const floatInVec &y );
\r
437 // Set the z element of a 3-D vector (scalar data contained in vector data type)
\r
439 __forceinline Vector3 & setZ( const floatInVec &z );
\r
441 // Get the x element of a 3-D vector
\r
443 __forceinline const floatInVec getX( ) const;
\r
445 // Get the y element of a 3-D vector
\r
447 __forceinline const floatInVec getY( ) const;
\r
449 // Get the z element of a 3-D vector
\r
451 __forceinline const floatInVec getZ( ) const;
\r
453 // Set an x, y, or z element of a 3-D vector by index
\r
455 __forceinline Vector3 & setElem( int idx, float value );
\r
457 // Set an x, y, or z element of a 3-D vector by index (scalar data contained in vector data type)
\r
459 __forceinline Vector3 & setElem( int idx, const floatInVec &value );
\r
461 // Get an x, y, or z element of a 3-D vector by index
\r
463 __forceinline const floatInVec getElem( int idx ) const;
\r
465 // Subscripting operator to set or get an element
\r
467 __forceinline VecIdx operator []( int idx );
\r
469 // Subscripting operator to get an element
\r
471 __forceinline const floatInVec operator []( int idx ) const;
\r
473 // Add two 3-D vectors
\r
475 __forceinline const Vector3 operator +( const Vector3 &vec ) const;
\r
477 // Subtract a 3-D vector from another 3-D vector
\r
479 __forceinline const Vector3 operator -( const Vector3 &vec ) const;
\r
481 // Add a 3-D vector to a 3-D point
\r
483 __forceinline const Point3 operator +( const Point3 &pnt ) const;
\r
485 // Multiply a 3-D vector by a scalar
\r
487 __forceinline const Vector3 operator *( float scalar ) const;
\r
489 // Divide a 3-D vector by a scalar
\r
491 __forceinline const Vector3 operator /( float scalar ) const;
\r
493 // Multiply a 3-D vector by a scalar (scalar data contained in vector data type)
\r
495 __forceinline const Vector3 operator *( const floatInVec &scalar ) const;
\r
497 // Divide a 3-D vector by a scalar (scalar data contained in vector data type)
\r
499 __forceinline const Vector3 operator /( const floatInVec &scalar ) const;
\r
501 // Perform compound assignment and addition with a 3-D vector
\r
503 __forceinline Vector3 & operator +=( const Vector3 &vec );
\r
505 // Perform compound assignment and subtraction by a 3-D vector
\r
507 __forceinline Vector3 & operator -=( const Vector3 &vec );
\r
509 // Perform compound assignment and multiplication by a scalar
\r
511 __forceinline Vector3 & operator *=( float scalar );
\r
513 // Perform compound assignment and division by a scalar
\r
515 __forceinline Vector3 & operator /=( float scalar );
\r
517 // Perform compound assignment and multiplication by a scalar (scalar data contained in vector data type)
\r
519 __forceinline Vector3 & operator *=( const floatInVec &scalar );
\r
521 // Perform compound assignment and division by a scalar (scalar data contained in vector data type)
\r
523 __forceinline Vector3 & operator /=( const floatInVec &scalar );
\r
525 // Negate all elements of a 3-D vector
\r
527 __forceinline const Vector3 operator -( ) const;
\r
529 // Construct x axis
\r
531 static __forceinline const Vector3 xAxis( );
\r
533 // Construct y axis
\r
535 static __forceinline const Vector3 yAxis( );
\r
537 // Construct z axis
\r
539 static __forceinline const Vector3 zAxis( );
\r
543 // Multiply a 3-D vector by a scalar
\r
545 __forceinline const Vector3 operator *( float scalar, const Vector3 &vec );
\r
547 // Multiply a 3-D vector by a scalar (scalar data contained in vector data type)
\r
549 __forceinline const Vector3 operator *( const floatInVec &scalar, const Vector3 &vec );
\r
551 // Multiply two 3-D vectors per element
\r
553 __forceinline const Vector3 mulPerElem( const Vector3 &vec0, const Vector3 &vec1 );
\r
555 // Divide two 3-D vectors per element
\r
557 // Floating-point behavior matches standard library function divf4.
\r
559 __forceinline const Vector3 divPerElem( const Vector3 &vec0, const Vector3 &vec1 );
\r
561 // Compute the reciprocal of a 3-D vector per element
\r
563 // Floating-point behavior matches standard library function recipf4.
\r
565 __forceinline const Vector3 recipPerElem( const Vector3 &vec );
\r
567 // Compute the absolute value of a 3-D vector per element
\r
569 __forceinline const Vector3 absPerElem( const Vector3 &vec );
\r
571 // Copy sign from one 3-D vector to another, per element
\r
573 __forceinline const Vector3 copySignPerElem( const Vector3 &vec0, const Vector3 &vec1 );
\r
575 // Maximum of two 3-D vectors per element
\r
577 __forceinline const Vector3 maxPerElem( const Vector3 &vec0, const Vector3 &vec1 );
\r
579 // Minimum of two 3-D vectors per element
\r
581 __forceinline const Vector3 minPerElem( const Vector3 &vec0, const Vector3 &vec1 );
\r
583 // Maximum element of a 3-D vector
\r
585 __forceinline const floatInVec maxElem( const Vector3 &vec );
\r
587 // Minimum element of a 3-D vector
\r
589 __forceinline const floatInVec minElem( const Vector3 &vec );
\r
591 // Compute the sum of all elements of a 3-D vector
\r
593 __forceinline const floatInVec sum( const Vector3 &vec );
\r
595 // Compute the dot product of two 3-D vectors
\r
597 __forceinline const floatInVec dot( const Vector3 &vec0, const Vector3 &vec1 );
\r
599 // Compute the square of the length of a 3-D vector
\r
601 __forceinline const floatInVec lengthSqr( const Vector3 &vec );
\r
603 // Compute the length of a 3-D vector
\r
605 __forceinline const floatInVec length( const Vector3 &vec );
\r
607 // Normalize a 3-D vector
\r
609 // The result is unpredictable when all elements of vec are at or near zero.
\r
611 __forceinline const Vector3 normalize( const Vector3 &vec );
\r
613 // Compute cross product of two 3-D vectors
\r
615 __forceinline const Vector3 cross( const Vector3 &vec0, const Vector3 &vec1 );
\r
617 // Outer product of two 3-D vectors
\r
619 __forceinline const Matrix3 outer( const Vector3 &vec0, const Vector3 &vec1 );
\r
621 // Pre-multiply a row vector by a 3x3 matrix
\r
623 // Slower than column post-multiply.
\r
625 __forceinline const Vector3 rowMul( const Vector3 &vec, const Matrix3 & mat );
\r
627 // Cross-product matrix of a 3-D vector
\r
629 __forceinline const Matrix3 crossMatrix( const Vector3 &vec );
\r
631 // Create cross-product matrix and multiply
\r
633 // Faster than separately creating a cross-product matrix and multiplying.
\r
635 __forceinline const Matrix3 crossMatrixMul( const Vector3 &vec, const Matrix3 & mat );
\r
637 // Linear interpolation between two 3-D vectors
\r
639 // Does not clamp t between 0 and 1.
\r
641 __forceinline const Vector3 lerp( float t, const Vector3 &vec0, const Vector3 &vec1 );
\r
643 // Linear interpolation between two 3-D vectors (scalar data contained in vector data type)
\r
645 // Does not clamp t between 0 and 1.
\r
647 __forceinline const Vector3 lerp( const floatInVec &t, const Vector3 &vec0, const Vector3 &vec1 );
\r
649 // Spherical linear interpolation between two 3-D vectors
\r
651 // The result is unpredictable if the vectors point in opposite directions.
\r
652 // Does not clamp t between 0 and 1.
\r
654 __forceinline const Vector3 slerp( float t, const Vector3 &unitVec0, const Vector3 &unitVec1 );
\r
656 // Spherical linear interpolation between two 3-D vectors (scalar data contained in vector data type)
\r
658 // The result is unpredictable if the vectors point in opposite directions.
\r
659 // Does not clamp t between 0 and 1.
\r
661 __forceinline const Vector3 slerp( const floatInVec &t, const Vector3 &unitVec0, const Vector3 &unitVec1 );
\r
663 // Conditionally select between two 3-D vectors
\r
665 // This function uses a conditional select instruction to avoid a branch.
\r
666 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
667 // Use the boolInVec version for better performance.
\r
669 __forceinline const Vector3 select( const Vector3 &vec0, const Vector3 &vec1, bool select1 );
\r
671 // Conditionally select between two 3-D vectors (scalar data contained in vector data type)
\r
673 // This function uses a conditional select instruction to avoid a branch.
\r
675 __forceinline const Vector3 select( const Vector3 &vec0, const Vector3 &vec1, const boolInVec &select1 );
\r
677 // Store x, y, and z elements of 3-D vector in first three words of a quadword, preserving fourth word
\r
679 __forceinline void storeXYZ( const Vector3 &vec, __m128 * quad );
\r
681 // Load four three-float 3-D vectors, stored in three quadwords
\r
683 __forceinline void loadXYZArray( Vector3 & vec0, Vector3 & vec1, Vector3 & vec2, Vector3 & vec3, const __m128 * threeQuads );
\r
685 // Store four 3-D vectors in three quadwords
\r
687 __forceinline void storeXYZArray( const Vector3 &vec0, const Vector3 &vec1, const Vector3 &vec2, const Vector3 &vec3, __m128 * threeQuads );
\r
689 // Store eight 3-D vectors as half-floats
\r
691 __forceinline void storeHalfFloats( const Vector3 &vec0, const Vector3 &vec1, const Vector3 &vec2, const Vector3 &vec3, const Vector3 &vec4, const Vector3 &vec5, const Vector3 &vec6, const Vector3 &vec7, vec_ushort8 * threeQuads );
\r
693 #ifdef _VECTORMATH_DEBUG
\r
695 // Print a 3-D vector
\r
697 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
699 __forceinline void print( const Vector3 &vec );
\r
701 // Print a 3-D vector and an associated string identifier
\r
703 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
705 __forceinline void print( const Vector3 &vec, const char * name );
\r
709 // A 4-D vector in array-of-structures format
\r
716 // Default constructor; does no initialization
\r
718 __forceinline Vector4( ) { };
\r
720 // Construct a 4-D vector from x, y, z, and w elements
\r
722 __forceinline Vector4( float x, float y, float z, float w );
\r
724 // Construct a 4-D vector from x, y, z, and w elements (scalar data contained in vector data type)
\r
726 __forceinline Vector4( const floatInVec &x, const floatInVec &y, const floatInVec &z, const floatInVec &w );
\r
728 // Construct a 4-D vector from a 3-D vector and a scalar
\r
730 __forceinline Vector4( const Vector3 &xyz, float w );
\r
732 // Construct a 4-D vector from a 3-D vector and a scalar (scalar data contained in vector data type)
\r
734 __forceinline Vector4( const Vector3 &xyz, const floatInVec &w );
\r
736 // Copy x, y, and z from a 3-D vector into a 4-D vector, and set w to 0
\r
738 explicit __forceinline Vector4( const Vector3 &vec );
\r
740 // Copy x, y, and z from a 3-D point into a 4-D vector, and set w to 1
\r
742 explicit __forceinline Vector4( const Point3 &pnt );
\r
744 // Copy elements from a quaternion into a 4-D vector
\r
746 explicit __forceinline Vector4( const Quat &quat );
\r
748 // Set all elements of a 4-D vector to the same scalar value
\r
750 explicit __forceinline Vector4( float scalar );
\r
752 // Set all elements of a 4-D vector to the same scalar value (scalar data contained in vector data type)
\r
754 explicit __forceinline Vector4( const floatInVec &scalar );
\r
756 // Set vector float data in a 4-D vector
\r
758 explicit __forceinline Vector4( __m128 vf4 );
\r
760 // Get vector float data from a 4-D vector
\r
762 __forceinline __m128 get128( ) const;
\r
764 // Assign one 4-D vector to another
\r
766 __forceinline Vector4 & operator =( const Vector4 &vec );
\r
768 // Set the x, y, and z elements of a 4-D vector
\r
770 // This function does not change the w element.
\r
772 __forceinline Vector4 & setXYZ( const Vector3 &vec );
\r
774 // Get the x, y, and z elements of a 4-D vector
\r
776 __forceinline const Vector3 getXYZ( ) const;
\r
778 // Set the x element of a 4-D vector
\r
780 __forceinline Vector4 & setX( float x );
\r
782 // Set the y element of a 4-D vector
\r
784 __forceinline Vector4 & setY( float y );
\r
786 // Set the z element of a 4-D vector
\r
788 __forceinline Vector4 & setZ( float z );
\r
790 // Set the w element of a 4-D vector
\r
792 __forceinline Vector4 & setW( float w );
\r
794 // Set the x element of a 4-D vector (scalar data contained in vector data type)
\r
796 __forceinline Vector4 & setX( const floatInVec &x );
\r
798 // Set the y element of a 4-D vector (scalar data contained in vector data type)
\r
800 __forceinline Vector4 & setY( const floatInVec &y );
\r
802 // Set the z element of a 4-D vector (scalar data contained in vector data type)
\r
804 __forceinline Vector4 & setZ( const floatInVec &z );
\r
806 // Set the w element of a 4-D vector (scalar data contained in vector data type)
\r
808 __forceinline Vector4 & setW( const floatInVec &w );
\r
810 // Get the x element of a 4-D vector
\r
812 __forceinline const floatInVec getX( ) const;
\r
814 // Get the y element of a 4-D vector
\r
816 __forceinline const floatInVec getY( ) const;
\r
818 // Get the z element of a 4-D vector
\r
820 __forceinline const floatInVec getZ( ) const;
\r
822 // Get the w element of a 4-D vector
\r
824 __forceinline const floatInVec getW( ) const;
\r
826 // Set an x, y, z, or w element of a 4-D vector by index
\r
828 __forceinline Vector4 & setElem( int idx, float value );
\r
830 // Set an x, y, z, or w element of a 4-D vector by index (scalar data contained in vector data type)
\r
832 __forceinline Vector4 & setElem( int idx, const floatInVec &value );
\r
834 // Get an x, y, z, or w element of a 4-D vector by index
\r
836 __forceinline const floatInVec getElem( int idx ) const;
\r
838 // Subscripting operator to set or get an element
\r
840 __forceinline VecIdx operator []( int idx );
\r
842 // Subscripting operator to get an element
\r
844 __forceinline const floatInVec operator []( int idx ) const;
\r
846 // Add two 4-D vectors
\r
848 __forceinline const Vector4 operator +( const Vector4 &vec ) const;
\r
850 // Subtract a 4-D vector from another 4-D vector
\r
852 __forceinline const Vector4 operator -( const Vector4 &vec ) const;
\r
854 // Multiply a 4-D vector by a scalar
\r
856 __forceinline const Vector4 operator *( float scalar ) const;
\r
858 // Divide a 4-D vector by a scalar
\r
860 __forceinline const Vector4 operator /( float scalar ) const;
\r
862 // Multiply a 4-D vector by a scalar (scalar data contained in vector data type)
\r
864 __forceinline const Vector4 operator *( const floatInVec &scalar ) const;
\r
866 // Divide a 4-D vector by a scalar (scalar data contained in vector data type)
\r
868 __forceinline const Vector4 operator /( const floatInVec &scalar ) const;
\r
870 // Perform compound assignment and addition with a 4-D vector
\r
872 __forceinline Vector4 & operator +=( const Vector4 &vec );
\r
874 // Perform compound assignment and subtraction by a 4-D vector
\r
876 __forceinline Vector4 & operator -=( const Vector4 &vec );
\r
878 // Perform compound assignment and multiplication by a scalar
\r
880 __forceinline Vector4 & operator *=( float scalar );
\r
882 // Perform compound assignment and division by a scalar
\r
884 __forceinline Vector4 & operator /=( float scalar );
\r
886 // Perform compound assignment and multiplication by a scalar (scalar data contained in vector data type)
\r
888 __forceinline Vector4 & operator *=( const floatInVec &scalar );
\r
890 // Perform compound assignment and division by a scalar (scalar data contained in vector data type)
\r
892 __forceinline Vector4 & operator /=( const floatInVec &scalar );
\r
894 // Negate all elements of a 4-D vector
\r
896 __forceinline const Vector4 operator -( ) const;
\r
898 // Construct x axis
\r
900 static __forceinline const Vector4 xAxis( );
\r
902 // Construct y axis
\r
904 static __forceinline const Vector4 yAxis( );
\r
906 // Construct z axis
\r
908 static __forceinline const Vector4 zAxis( );
\r
910 // Construct w axis
\r
912 static __forceinline const Vector4 wAxis( );
\r
916 // Multiply a 4-D vector by a scalar
\r
918 __forceinline const Vector4 operator *( float scalar, const Vector4 &vec );
\r
920 // Multiply a 4-D vector by a scalar (scalar data contained in vector data type)
\r
922 __forceinline const Vector4 operator *( const floatInVec &scalar, const Vector4 &vec );
\r
924 // Multiply two 4-D vectors per element
\r
926 __forceinline const Vector4 mulPerElem( const Vector4 &vec0, const Vector4 &vec1 );
\r
928 // Divide two 4-D vectors per element
\r
930 // Floating-point behavior matches standard library function divf4.
\r
932 __forceinline const Vector4 divPerElem( const Vector4 &vec0, const Vector4 &vec1 );
\r
934 // Compute the reciprocal of a 4-D vector per element
\r
936 // Floating-point behavior matches standard library function recipf4.
\r
938 __forceinline const Vector4 recipPerElem( const Vector4 &vec );
\r
940 // Compute the absolute value of a 4-D vector per element
\r
942 __forceinline const Vector4 absPerElem( const Vector4 &vec );
\r
944 // Copy sign from one 4-D vector to another, per element
\r
946 __forceinline const Vector4 copySignPerElem( const Vector4 &vec0, const Vector4 &vec1 );
\r
948 // Maximum of two 4-D vectors per element
\r
950 __forceinline const Vector4 maxPerElem( const Vector4 &vec0, const Vector4 &vec1 );
\r
952 // Minimum of two 4-D vectors per element
\r
954 __forceinline const Vector4 minPerElem( const Vector4 &vec0, const Vector4 &vec1 );
\r
956 // Maximum element of a 4-D vector
\r
958 __forceinline const floatInVec maxElem( const Vector4 &vec );
\r
960 // Minimum element of a 4-D vector
\r
962 __forceinline const floatInVec minElem( const Vector4 &vec );
\r
964 // Compute the sum of all elements of a 4-D vector
\r
966 __forceinline const floatInVec sum( const Vector4 &vec );
\r
968 // Compute the dot product of two 4-D vectors
\r
970 __forceinline const floatInVec dot( const Vector4 &vec0, const Vector4 &vec1 );
\r
972 // Compute the square of the length of a 4-D vector
\r
974 __forceinline const floatInVec lengthSqr( const Vector4 &vec );
\r
976 // Compute the length of a 4-D vector
\r
978 __forceinline const floatInVec length( const Vector4 &vec );
\r
980 // Normalize a 4-D vector
\r
982 // The result is unpredictable when all elements of vec are at or near zero.
\r
984 __forceinline const Vector4 normalize( const Vector4 &vec );
\r
986 // Outer product of two 4-D vectors
\r
988 __forceinline const Matrix4 outer( const Vector4 &vec0, const Vector4 &vec1 );
\r
990 // Linear interpolation between two 4-D vectors
\r
992 // Does not clamp t between 0 and 1.
\r
994 __forceinline const Vector4 lerp( float t, const Vector4 &vec0, const Vector4 &vec1 );
\r
996 // Linear interpolation between two 4-D vectors (scalar data contained in vector data type)
\r
998 // Does not clamp t between 0 and 1.
\r
1000 __forceinline const Vector4 lerp( const floatInVec &t, const Vector4 &vec0, const Vector4 &vec1 );
\r
1002 // Spherical linear interpolation between two 4-D vectors
\r
1004 // The result is unpredictable if the vectors point in opposite directions.
\r
1005 // Does not clamp t between 0 and 1.
\r
1007 __forceinline const Vector4 slerp( float t, const Vector4 &unitVec0, const Vector4 &unitVec1 );
\r
1009 // Spherical linear interpolation between two 4-D vectors (scalar data contained in vector data type)
\r
1011 // The result is unpredictable if the vectors point in opposite directions.
\r
1012 // Does not clamp t between 0 and 1.
\r
1014 __forceinline const Vector4 slerp( const floatInVec &t, const Vector4 &unitVec0, const Vector4 &unitVec1 );
\r
1016 // Conditionally select between two 4-D vectors
\r
1018 // This function uses a conditional select instruction to avoid a branch.
\r
1019 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
1020 // Use the boolInVec version for better performance.
\r
1022 __forceinline const Vector4 select( const Vector4 &vec0, const Vector4 &vec1, bool select1 );
\r
1024 // Conditionally select between two 4-D vectors (scalar data contained in vector data type)
\r
1026 // This function uses a conditional select instruction to avoid a branch.
\r
1028 __forceinline const Vector4 select( const Vector4 &vec0, const Vector4 &vec1, const boolInVec &select1 );
\r
1030 // Store four 4-D vectors as half-floats
\r
1032 __forceinline void storeHalfFloats( const Vector4 &vec0, const Vector4 &vec1, const Vector4 &vec2, const Vector4 &vec3, vec_ushort8 * twoQuads );
\r
1034 #ifdef _VECTORMATH_DEBUG
\r
1036 // Print a 4-D vector
\r
1038 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1040 __forceinline void print( const Vector4 &vec );
\r
1042 // Print a 4-D vector and an associated string identifier
\r
1044 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1046 __forceinline void print( const Vector4 &vec, const char * name );
\r
1050 // A 3-D point in array-of-structures format
\r
1057 // Default constructor; does no initialization
\r
1059 __forceinline Point3( ) { };
\r
1061 // Construct a 3-D point from x, y, and z elements
\r
1063 __forceinline Point3( float x, float y, float z );
\r
1065 // Construct a 3-D point from x, y, and z elements (scalar data contained in vector data type)
\r
1067 __forceinline Point3( const floatInVec &x, const floatInVec &y, const floatInVec &z );
\r
1069 // Copy elements from a 3-D vector into a 3-D point
\r
1071 explicit __forceinline Point3( const Vector3 &vec );
\r
1073 // Set all elements of a 3-D point to the same scalar value
\r
1075 explicit __forceinline Point3( float scalar );
\r
1077 // Set all elements of a 3-D point to the same scalar value (scalar data contained in vector data type)
\r
1079 explicit __forceinline Point3( const floatInVec &scalar );
\r
1081 // Set vector float data in a 3-D point
\r
1083 explicit __forceinline Point3( __m128 vf4 );
\r
1085 // Get vector float data from a 3-D point
\r
1087 __forceinline __m128 get128( ) const;
\r
1089 // Assign one 3-D point to another
\r
1091 __forceinline Point3 & operator =( const Point3 &pnt );
\r
1093 // Set the x element of a 3-D point
\r
1095 __forceinline Point3 & setX( float x );
\r
1097 // Set the y element of a 3-D point
\r
1099 __forceinline Point3 & setY( float y );
\r
1101 // Set the z element of a 3-D point
\r
1103 __forceinline Point3 & setZ( float z );
\r
1105 // Set the x element of a 3-D point (scalar data contained in vector data type)
\r
1107 __forceinline Point3 & setX( const floatInVec &x );
\r
1109 // Set the y element of a 3-D point (scalar data contained in vector data type)
\r
1111 __forceinline Point3 & setY( const floatInVec &y );
\r
1113 // Set the z element of a 3-D point (scalar data contained in vector data type)
\r
1115 __forceinline Point3 & setZ( const floatInVec &z );
\r
1117 // Get the x element of a 3-D point
\r
1119 __forceinline const floatInVec getX( ) const;
\r
1121 // Get the y element of a 3-D point
\r
1123 __forceinline const floatInVec getY( ) const;
\r
1125 // Get the z element of a 3-D point
\r
1127 __forceinline const floatInVec getZ( ) const;
\r
1129 // Set an x, y, or z element of a 3-D point by index
\r
1131 __forceinline Point3 & setElem( int idx, float value );
\r
1133 // Set an x, y, or z element of a 3-D point by index (scalar data contained in vector data type)
\r
1135 __forceinline Point3 & setElem( int idx, const floatInVec &value );
\r
1137 // Get an x, y, or z element of a 3-D point by index
\r
1139 __forceinline const floatInVec getElem( int idx ) const;
\r
1141 // Subscripting operator to set or get an element
\r
1143 __forceinline VecIdx operator []( int idx );
\r
1145 // Subscripting operator to get an element
\r
1147 __forceinline const floatInVec operator []( int idx ) const;
\r
1149 // Subtract a 3-D point from another 3-D point
\r
1151 __forceinline const Vector3 operator -( const Point3 &pnt ) const;
\r
1153 // Add a 3-D point to a 3-D vector
\r
1155 __forceinline const Point3 operator +( const Vector3 &vec ) const;
\r
1157 // Subtract a 3-D vector from a 3-D point
\r
1159 __forceinline const Point3 operator -( const Vector3 &vec ) const;
\r
1161 // Perform compound assignment and addition with a 3-D vector
\r
1163 __forceinline Point3 & operator +=( const Vector3 &vec );
\r
1165 // Perform compound assignment and subtraction by a 3-D vector
\r
1167 __forceinline Point3 & operator -=( const Vector3 &vec );
\r
1171 // Multiply two 3-D points per element
\r
1173 __forceinline const Point3 mulPerElem( const Point3 &pnt0, const Point3 &pnt1 );
\r
1175 // Divide two 3-D points per element
\r
1177 // Floating-point behavior matches standard library function divf4.
\r
1179 __forceinline const Point3 divPerElem( const Point3 &pnt0, const Point3 &pnt1 );
\r
1181 // Compute the reciprocal of a 3-D point per element
\r
1183 // Floating-point behavior matches standard library function recipf4.
\r
1185 __forceinline const Point3 recipPerElem( const Point3 &pnt );
\r
1187 // Compute the absolute value of a 3-D point per element
\r
1189 __forceinline const Point3 absPerElem( const Point3 &pnt );
\r
1191 // Copy sign from one 3-D point to another, per element
\r
1193 __forceinline const Point3 copySignPerElem( const Point3 &pnt0, const Point3 &pnt1 );
\r
1195 // Maximum of two 3-D points per element
\r
1197 __forceinline const Point3 maxPerElem( const Point3 &pnt0, const Point3 &pnt1 );
\r
1199 // Minimum of two 3-D points per element
\r
1201 __forceinline const Point3 minPerElem( const Point3 &pnt0, const Point3 &pnt1 );
\r
1203 // Maximum element of a 3-D point
\r
1205 __forceinline const floatInVec maxElem( const Point3 &pnt );
\r
1207 // Minimum element of a 3-D point
\r
1209 __forceinline const floatInVec minElem( const Point3 &pnt );
\r
1211 // Compute the sum of all elements of a 3-D point
\r
1213 __forceinline const floatInVec sum( const Point3 &pnt );
\r
1215 // Apply uniform scale to a 3-D point
\r
1217 __forceinline const Point3 scale( const Point3 &pnt, float scaleVal );
\r
1219 // Apply uniform scale to a 3-D point (scalar data contained in vector data type)
\r
1221 __forceinline const Point3 scale( const Point3 &pnt, const floatInVec &scaleVal );
\r
1223 // Apply non-uniform scale to a 3-D point
\r
1225 __forceinline const Point3 scale( const Point3 &pnt, const Vector3 &scaleVec );
\r
1227 // Scalar projection of a 3-D point on a unit-length 3-D vector
\r
1229 __forceinline const floatInVec projection( const Point3 &pnt, const Vector3 &unitVec );
\r
1231 // Compute the square of the distance of a 3-D point from the coordinate-system origin
\r
1233 __forceinline const floatInVec distSqrFromOrigin( const Point3 &pnt );
\r
1235 // Compute the distance of a 3-D point from the coordinate-system origin
\r
1237 __forceinline const floatInVec distFromOrigin( const Point3 &pnt );
\r
1239 // Compute the square of the distance between two 3-D points
\r
1241 __forceinline const floatInVec distSqr( const Point3 &pnt0, const Point3 &pnt1 );
\r
1243 // Compute the distance between two 3-D points
\r
1245 __forceinline const floatInVec dist( const Point3 &pnt0, const Point3 &pnt1 );
\r
1247 // Linear interpolation between two 3-D points
\r
1249 // Does not clamp t between 0 and 1.
\r
1251 __forceinline const Point3 lerp( float t, const Point3 &pnt0, const Point3 &pnt1 );
\r
1253 // Linear interpolation between two 3-D points (scalar data contained in vector data type)
\r
1255 // Does not clamp t between 0 and 1.
\r
1257 __forceinline const Point3 lerp( const floatInVec &t, const Point3 &pnt0, const Point3 &pnt1 );
\r
1259 // Conditionally select between two 3-D points
\r
1261 // This function uses a conditional select instruction to avoid a branch.
\r
1262 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
1263 // Use the boolInVec version for better performance.
\r
1265 __forceinline const Point3 select( const Point3 &pnt0, const Point3 &pnt1, bool select1 );
\r
1267 // Conditionally select between two 3-D points (scalar data contained in vector data type)
\r
1269 // This function uses a conditional select instruction to avoid a branch.
\r
1271 __forceinline const Point3 select( const Point3 &pnt0, const Point3 &pnt1, const boolInVec &select1 );
\r
1273 // Store x, y, and z elements of 3-D point in first three words of a quadword, preserving fourth word
\r
1275 __forceinline void storeXYZ( const Point3 &pnt, __m128 * quad );
\r
1277 // Load four three-float 3-D points, stored in three quadwords
\r
1279 __forceinline void loadXYZArray( Point3 & pnt0, Point3 & pnt1, Point3 & pnt2, Point3 & pnt3, const __m128 * threeQuads );
\r
1281 // Store four 3-D points in three quadwords
\r
1283 __forceinline void storeXYZArray( const Point3 &pnt0, const Point3 &pnt1, const Point3 &pnt2, const Point3 &pnt3, __m128 * threeQuads );
\r
1285 // Store eight 3-D points as half-floats
\r
1287 __forceinline void storeHalfFloats( const Point3 &pnt0, const Point3 &pnt1, const Point3 &pnt2, const Point3 &pnt3, const Point3 &pnt4, const Point3 &pnt5, const Point3 &pnt6, const Point3 &pnt7, vec_ushort8 * threeQuads );
\r
1289 #ifdef _VECTORMATH_DEBUG
\r
1291 // Print a 3-D point
\r
1293 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1295 __forceinline void print( const Point3 &pnt );
\r
1297 // Print a 3-D point and an associated string identifier
\r
1299 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1301 __forceinline void print( const Point3 &pnt, const char * name );
\r
1305 // A quaternion in array-of-structures format
\r
1312 // Default constructor; does no initialization
\r
1314 __forceinline Quat( ) { };
\r
1316 __forceinline Quat::Quat(const Quat& quat);
\r
1318 // Construct a quaternion from x, y, z, and w elements
\r
1320 __forceinline Quat( float x, float y, float z, float w );
\r
1322 // Construct a quaternion from x, y, z, and w elements (scalar data contained in vector data type)
\r
1324 __forceinline Quat( const floatInVec &x, const floatInVec &y, const floatInVec &z, const floatInVec &w );
\r
1326 // Construct a quaternion from a 3-D vector and a scalar
\r
1328 __forceinline Quat( const Vector3 &xyz, float w );
\r
1330 // Construct a quaternion from a 3-D vector and a scalar (scalar data contained in vector data type)
\r
1332 __forceinline Quat( const Vector3 &xyz, const floatInVec &w );
\r
1334 // Copy elements from a 4-D vector into a quaternion
\r
1336 explicit __forceinline Quat( const Vector4 &vec );
\r
1338 // Convert a rotation matrix to a unit-length quaternion
\r
1340 explicit __forceinline Quat( const Matrix3 & rotMat );
\r
1342 // Set all elements of a quaternion to the same scalar value
\r
1344 explicit __forceinline Quat( float scalar );
\r
1346 // Set all elements of a quaternion to the same scalar value (scalar data contained in vector data type)
\r
1348 explicit __forceinline Quat( const floatInVec &scalar );
\r
1350 // Set vector float data in a quaternion
\r
1352 explicit __forceinline Quat( __m128 vf4 );
\r
1354 // Get vector float data from a quaternion
\r
1356 __forceinline __m128 get128( ) const;
\r
1358 // Set a quaterion from vector float data
\r
1360 __forceinline void set128(vec_float4 vec);
\r
1362 // Assign one quaternion to another
\r
1364 __forceinline Quat & operator =( const Quat &quat );
\r
1366 // Set the x, y, and z elements of a quaternion
\r
1368 // This function does not change the w element.
\r
1370 __forceinline Quat & setXYZ( const Vector3 &vec );
\r
1372 // Get the x, y, and z elements of a quaternion
\r
1374 __forceinline const Vector3 getXYZ( ) const;
\r
1376 // Set the x element of a quaternion
\r
1378 __forceinline Quat & setX( float x );
\r
1380 // Set the y element of a quaternion
\r
1382 __forceinline Quat & setY( float y );
\r
1384 // Set the z element of a quaternion
\r
1386 __forceinline Quat & setZ( float z );
\r
1388 // Set the w element of a quaternion
\r
1390 __forceinline Quat & setW( float w );
\r
1392 // Set the x element of a quaternion (scalar data contained in vector data type)
\r
1394 __forceinline Quat & setX( const floatInVec &x );
\r
1396 // Set the y element of a quaternion (scalar data contained in vector data type)
\r
1398 __forceinline Quat & setY( const floatInVec &y );
\r
1400 // Set the z element of a quaternion (scalar data contained in vector data type)
\r
1402 __forceinline Quat & setZ( const floatInVec &z );
\r
1404 // Set the w element of a quaternion (scalar data contained in vector data type)
\r
1406 __forceinline Quat & setW( const floatInVec &w );
\r
1408 // Get the x element of a quaternion
\r
1410 __forceinline const floatInVec getX( ) const;
\r
1412 // Get the y element of a quaternion
\r
1414 __forceinline const floatInVec getY( ) const;
\r
1416 // Get the z element of a quaternion
\r
1418 __forceinline const floatInVec getZ( ) const;
\r
1420 // Get the w element of a quaternion
\r
1422 __forceinline const floatInVec getW( ) const;
\r
1424 // Set an x, y, z, or w element of a quaternion by index
\r
1426 __forceinline Quat & setElem( int idx, float value );
\r
1428 // Set an x, y, z, or w element of a quaternion by index (scalar data contained in vector data type)
\r
1430 __forceinline Quat & setElem( int idx, const floatInVec &value );
\r
1432 // Get an x, y, z, or w element of a quaternion by index
\r
1434 __forceinline const floatInVec getElem( int idx ) const;
\r
1436 // Subscripting operator to set or get an element
\r
1438 __forceinline VecIdx operator []( int idx );
\r
1440 // Subscripting operator to get an element
\r
1442 __forceinline const floatInVec operator []( int idx ) const;
\r
1444 // Add two quaternions
\r
1446 __forceinline const Quat operator +( const Quat &quat ) const;
\r
1448 // Subtract a quaternion from another quaternion
\r
1450 __forceinline const Quat operator -( const Quat &quat ) const;
\r
1452 // Multiply two quaternions
\r
1454 __forceinline const Quat operator *( const Quat &quat ) const;
\r
1456 // Multiply a quaternion by a scalar
\r
1458 __forceinline const Quat operator *( float scalar ) const;
\r
1460 // Divide a quaternion by a scalar
\r
1462 __forceinline const Quat operator /( float scalar ) const;
\r
1464 // Multiply a quaternion by a scalar (scalar data contained in vector data type)
\r
1466 __forceinline const Quat operator *( const floatInVec &scalar ) const;
\r
1468 // Divide a quaternion by a scalar (scalar data contained in vector data type)
\r
1470 __forceinline const Quat operator /( const floatInVec &scalar ) const;
\r
1472 // Perform compound assignment and addition with a quaternion
\r
1474 __forceinline Quat & operator +=( const Quat &quat );
\r
1476 // Perform compound assignment and subtraction by a quaternion
\r
1478 __forceinline Quat & operator -=( const Quat &quat );
\r
1480 // Perform compound assignment and multiplication by a quaternion
\r
1482 __forceinline Quat & operator *=( const Quat &quat );
\r
1484 // Perform compound assignment and multiplication by a scalar
\r
1486 __forceinline Quat & operator *=( float scalar );
\r
1488 // Perform compound assignment and division by a scalar
\r
1490 __forceinline Quat & operator /=( float scalar );
\r
1492 // Perform compound assignment and multiplication by a scalar (scalar data contained in vector data type)
\r
1494 __forceinline Quat & operator *=( const floatInVec &scalar );
\r
1496 // Perform compound assignment and division by a scalar (scalar data contained in vector data type)
\r
1498 __forceinline Quat & operator /=( const floatInVec &scalar );
\r
1500 // Negate all elements of a quaternion
\r
1502 __forceinline const Quat operator -( ) const;
\r
1504 // Construct an identity quaternion
\r
1506 static __forceinline const Quat identity( );
\r
1508 // Construct a quaternion to rotate between two unit-length 3-D vectors
\r
1510 // The result is unpredictable if unitVec0 and unitVec1 point in opposite directions.
\r
1512 static __forceinline const Quat rotation( const Vector3 &unitVec0, const Vector3 &unitVec1 );
\r
1514 // Construct a quaternion to rotate around a unit-length 3-D vector
\r
1516 static __forceinline const Quat rotation( float radians, const Vector3 &unitVec );
\r
1518 // Construct a quaternion to rotate around a unit-length 3-D vector (scalar data contained in vector data type)
\r
1520 static __forceinline const Quat rotation( const floatInVec &radians, const Vector3 &unitVec );
\r
1522 // Construct a quaternion to rotate around the x axis
\r
1524 static __forceinline const Quat rotationX( float radians );
\r
1526 // Construct a quaternion to rotate around the y axis
\r
1528 static __forceinline const Quat rotationY( float radians );
\r
1530 // Construct a quaternion to rotate around the z axis
\r
1532 static __forceinline const Quat rotationZ( float radians );
\r
1534 // Construct a quaternion to rotate around the x axis (scalar data contained in vector data type)
\r
1536 static __forceinline const Quat rotationX( const floatInVec &radians );
\r
1538 // Construct a quaternion to rotate around the y axis (scalar data contained in vector data type)
\r
1540 static __forceinline const Quat rotationY( const floatInVec &radians );
\r
1542 // Construct a quaternion to rotate around the z axis (scalar data contained in vector data type)
\r
1544 static __forceinline const Quat rotationZ( const floatInVec &radians );
\r
1548 // Multiply a quaternion by a scalar
\r
1550 __forceinline const Quat operator *( float scalar, const Quat &quat );
\r
1552 // Multiply a quaternion by a scalar (scalar data contained in vector data type)
\r
1554 __forceinline const Quat operator *( const floatInVec &scalar, const Quat &quat );
\r
1556 // Compute the conjugate of a quaternion
\r
1558 __forceinline const Quat conj( const Quat &quat );
\r
1560 // Use a unit-length quaternion to rotate a 3-D vector
\r
1562 __forceinline const Vector3 rotate( const Quat &unitQuat, const Vector3 &vec );
\r
1564 // Compute the dot product of two quaternions
\r
1566 __forceinline const floatInVec dot( const Quat &quat0, const Quat &quat1 );
\r
1568 // Compute the norm of a quaternion
\r
1570 __forceinline const floatInVec norm( const Quat &quat );
\r
1572 // Compute the length of a quaternion
\r
1574 __forceinline const floatInVec length( const Quat &quat );
\r
1576 // Normalize a quaternion
\r
1578 // The result is unpredictable when all elements of quat are at or near zero.
\r
1580 __forceinline const Quat normalize( const Quat &quat );
\r
1582 // Linear interpolation between two quaternions
\r
1584 // Does not clamp t between 0 and 1.
\r
1586 __forceinline const Quat lerp( float t, const Quat &quat0, const Quat &quat1 );
\r
1588 // Linear interpolation between two quaternions (scalar data contained in vector data type)
\r
1590 // Does not clamp t between 0 and 1.
\r
1592 __forceinline const Quat lerp( const floatInVec &t, const Quat &quat0, const Quat &quat1 );
\r
1594 // Spherical linear interpolation between two quaternions
\r
1596 // Interpolates along the shortest path between orientations.
\r
1597 // Does not clamp t between 0 and 1.
\r
1599 __forceinline const Quat slerp( float t, const Quat &unitQuat0, const Quat &unitQuat1 );
\r
1601 // Spherical linear interpolation between two quaternions (scalar data contained in vector data type)
\r
1603 // Interpolates along the shortest path between orientations.
\r
1604 // Does not clamp t between 0 and 1.
\r
1606 __forceinline const Quat slerp( const floatInVec &t, const Quat &unitQuat0, const Quat &unitQuat1 );
\r
1608 // Spherical quadrangle interpolation
\r
1610 __forceinline const Quat squad( float t, const Quat &unitQuat0, const Quat &unitQuat1, const Quat &unitQuat2, const Quat &unitQuat3 );
\r
1612 // Spherical quadrangle interpolation (scalar data contained in vector data type)
\r
1614 __forceinline const Quat squad( const floatInVec &t, const Quat &unitQuat0, const Quat &unitQuat1, const Quat &unitQuat2, const Quat &unitQuat3 );
\r
1616 // Conditionally select between two quaternions
\r
1618 // This function uses a conditional select instruction to avoid a branch.
\r
1619 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
1620 // Use the boolInVec version for better performance.
\r
1622 __forceinline const Quat select( const Quat &quat0, const Quat &quat1, bool select1 );
\r
1624 // Conditionally select between two quaternions (scalar data contained in vector data type)
\r
1626 // This function uses a conditional select instruction to avoid a branch.
\r
1628 __forceinline const Quat select( const Quat &quat0, const Quat &quat1, const boolInVec &select1 );
\r
1630 #ifdef _VECTORMATH_DEBUG
\r
1632 // Print a quaternion
\r
1634 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1636 __forceinline void print( const Quat &quat );
\r
1638 // Print a quaternion and an associated string identifier
\r
1640 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1642 __forceinline void print( const Quat &quat, const char * name );
\r
1646 // A 3x3 matrix in array-of-structures format
\r
1655 // Default constructor; does no initialization
\r
1657 __forceinline Matrix3( ) { };
\r
1659 // Copy a 3x3 matrix
\r
1661 __forceinline Matrix3( const Matrix3 & mat );
\r
1663 // Construct a 3x3 matrix containing the specified columns
\r
1665 __forceinline Matrix3( const Vector3 &col0, const Vector3 &col1, const Vector3 &col2 );
\r
1667 // Construct a 3x3 rotation matrix from a unit-length quaternion
\r
1669 explicit __forceinline Matrix3( const Quat &unitQuat );
\r
1671 // Set all elements of a 3x3 matrix to the same scalar value
\r
1673 explicit __forceinline Matrix3( float scalar );
\r
1675 // Set all elements of a 3x3 matrix to the same scalar value (scalar data contained in vector data type)
\r
1677 explicit __forceinline Matrix3( const floatInVec &scalar );
\r
1679 // Assign one 3x3 matrix to another
\r
1681 __forceinline Matrix3 & operator =( const Matrix3 & mat );
\r
1683 // Set column 0 of a 3x3 matrix
\r
1685 __forceinline Matrix3 & setCol0( const Vector3 &col0 );
\r
1687 // Set column 1 of a 3x3 matrix
\r
1689 __forceinline Matrix3 & setCol1( const Vector3 &col1 );
\r
1691 // Set column 2 of a 3x3 matrix
\r
1693 __forceinline Matrix3 & setCol2( const Vector3 &col2 );
\r
1695 // Get column 0 of a 3x3 matrix
\r
1697 __forceinline const Vector3 getCol0( ) const;
\r
1699 // Get column 1 of a 3x3 matrix
\r
1701 __forceinline const Vector3 getCol1( ) const;
\r
1703 // Get column 2 of a 3x3 matrix
\r
1705 __forceinline const Vector3 getCol2( ) const;
\r
1707 // Set the column of a 3x3 matrix referred to by the specified index
\r
1709 __forceinline Matrix3 & setCol( int col, const Vector3 &vec );
\r
1711 // Set the row of a 3x3 matrix referred to by the specified index
\r
1713 __forceinline Matrix3 & setRow( int row, const Vector3 &vec );
\r
1715 // Get the column of a 3x3 matrix referred to by the specified index
\r
1717 __forceinline const Vector3 getCol( int col ) const;
\r
1719 // Get the row of a 3x3 matrix referred to by the specified index
\r
1721 __forceinline const Vector3 getRow( int row ) const;
\r
1723 // Subscripting operator to set or get a column
\r
1725 __forceinline Vector3 & operator []( int col );
\r
1727 // Subscripting operator to get a column
\r
1729 __forceinline const Vector3 operator []( int col ) const;
\r
1731 // Set the element of a 3x3 matrix referred to by column and row indices
\r
1733 __forceinline Matrix3 & setElem( int col, int row, float val );
\r
1735 // Set the element of a 3x3 matrix referred to by column and row indices (scalar data contained in vector data type)
\r
1737 __forceinline Matrix3 & setElem( int col, int row, const floatInVec &val );
\r
1739 // Get the element of a 3x3 matrix referred to by column and row indices
\r
1741 __forceinline const floatInVec getElem( int col, int row ) const;
\r
1743 // Add two 3x3 matrices
\r
1745 __forceinline const Matrix3 operator +( const Matrix3 & mat ) const;
\r
1747 // Subtract a 3x3 matrix from another 3x3 matrix
\r
1749 __forceinline const Matrix3 operator -( const Matrix3 & mat ) const;
\r
1751 // Negate all elements of a 3x3 matrix
\r
1753 __forceinline const Matrix3 operator -( ) const;
\r
1755 // Multiply a 3x3 matrix by a scalar
\r
1757 __forceinline const Matrix3 operator *( float scalar ) const;
\r
1759 // Multiply a 3x3 matrix by a scalar (scalar data contained in vector data type)
\r
1761 __forceinline const Matrix3 operator *( const floatInVec &scalar ) const;
\r
1763 // Multiply a 3x3 matrix by a 3-D vector
\r
1765 __forceinline const Vector3 operator *( const Vector3 &vec ) const;
\r
1767 // Multiply two 3x3 matrices
\r
1769 __forceinline const Matrix3 operator *( const Matrix3 & mat ) const;
\r
1771 // Perform compound assignment and addition with a 3x3 matrix
\r
1773 __forceinline Matrix3 & operator +=( const Matrix3 & mat );
\r
1775 // Perform compound assignment and subtraction by a 3x3 matrix
\r
1777 __forceinline Matrix3 & operator -=( const Matrix3 & mat );
\r
1779 // Perform compound assignment and multiplication by a scalar
\r
1781 __forceinline Matrix3 & operator *=( float scalar );
\r
1783 // Perform compound assignment and multiplication by a scalar (scalar data contained in vector data type)
\r
1785 __forceinline Matrix3 & operator *=( const floatInVec &scalar );
\r
1787 // Perform compound assignment and multiplication by a 3x3 matrix
\r
1789 __forceinline Matrix3 & operator *=( const Matrix3 & mat );
\r
1791 // Construct an identity 3x3 matrix
\r
1793 static __forceinline const Matrix3 identity( );
\r
1795 // Construct a 3x3 matrix to rotate around the x axis
\r
1797 static __forceinline const Matrix3 rotationX( float radians );
\r
1799 // Construct a 3x3 matrix to rotate around the y axis
\r
1801 static __forceinline const Matrix3 rotationY( float radians );
\r
1803 // Construct a 3x3 matrix to rotate around the z axis
\r
1805 static __forceinline const Matrix3 rotationZ( float radians );
\r
1807 // Construct a 3x3 matrix to rotate around the x axis (scalar data contained in vector data type)
\r
1809 static __forceinline const Matrix3 rotationX( const floatInVec &radians );
\r
1811 // Construct a 3x3 matrix to rotate around the y axis (scalar data contained in vector data type)
\r
1813 static __forceinline const Matrix3 rotationY( const floatInVec &radians );
\r
1815 // Construct a 3x3 matrix to rotate around the z axis (scalar data contained in vector data type)
\r
1817 static __forceinline const Matrix3 rotationZ( const floatInVec &radians );
\r
1819 // Construct a 3x3 matrix to rotate around the x, y, and z axes
\r
1821 static __forceinline const Matrix3 rotationZYX( const Vector3 &radiansXYZ );
\r
1823 // Construct a 3x3 matrix to rotate around a unit-length 3-D vector
\r
1825 static __forceinline const Matrix3 rotation( float radians, const Vector3 &unitVec );
\r
1827 // Construct a 3x3 matrix to rotate around a unit-length 3-D vector (scalar data contained in vector data type)
\r
1829 static __forceinline const Matrix3 rotation( const floatInVec &radians, const Vector3 &unitVec );
\r
1831 // Construct a rotation matrix from a unit-length quaternion
\r
1833 static __forceinline const Matrix3 rotation( const Quat &unitQuat );
\r
1835 // Construct a 3x3 matrix to perform scaling
\r
1837 static __forceinline const Matrix3 scale( const Vector3 &scaleVec );
\r
1840 // Multiply a 3x3 matrix by a scalar
\r
1842 __forceinline const Matrix3 operator *( float scalar, const Matrix3 & mat );
\r
1844 // Multiply a 3x3 matrix by a scalar (scalar data contained in vector data type)
\r
1846 __forceinline const Matrix3 operator *( const floatInVec &scalar, const Matrix3 & mat );
\r
1848 // Append (post-multiply) a scale transformation to a 3x3 matrix
\r
1850 // Faster than creating and multiplying a scale transformation matrix.
\r
1852 __forceinline const Matrix3 appendScale( const Matrix3 & mat, const Vector3 &scaleVec );
\r
1854 // Prepend (pre-multiply) a scale transformation to a 3x3 matrix
\r
1856 // Faster than creating and multiplying a scale transformation matrix.
\r
1858 __forceinline const Matrix3 prependScale( const Vector3 &scaleVec, const Matrix3 & mat );
\r
1860 // Multiply two 3x3 matrices per element
\r
1862 __forceinline const Matrix3 mulPerElem( const Matrix3 & mat0, const Matrix3 & mat1 );
\r
1864 // Compute the absolute value of a 3x3 matrix per element
\r
1866 __forceinline const Matrix3 absPerElem( const Matrix3 & mat );
\r
1868 // Transpose of a 3x3 matrix
\r
1870 __forceinline const Matrix3 transpose( const Matrix3 & mat );
\r
1872 // Compute the inverse of a 3x3 matrix
\r
1874 // Result is unpredictable when the determinant of mat is equal to or near 0.
\r
1876 __forceinline const Matrix3 inverse( const Matrix3 & mat );
\r
1878 // Determinant of a 3x3 matrix
\r
1880 __forceinline const floatInVec determinant( const Matrix3 & mat );
\r
1882 // Conditionally select between two 3x3 matrices
\r
1884 // This function uses a conditional select instruction to avoid a branch.
\r
1885 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
1886 // Use the boolInVec version for better performance.
\r
1888 __forceinline const Matrix3 select( const Matrix3 & mat0, const Matrix3 & mat1, bool select1 );
\r
1890 // Conditionally select between two 3x3 matrices (scalar data contained in vector data type)
\r
1892 // This function uses a conditional select instruction to avoid a branch.
\r
1894 __forceinline const Matrix3 select( const Matrix3 & mat0, const Matrix3 & mat1, const boolInVec &select1 );
\r
1896 #ifdef _VECTORMATH_DEBUG
\r
1898 // Print a 3x3 matrix
\r
1900 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1902 __forceinline void print( const Matrix3 & mat );
\r
1904 // Print a 3x3 matrix and an associated string identifier
\r
1906 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
1908 __forceinline void print( const Matrix3 & mat, const char * name );
\r
1912 // A 4x4 matrix in array-of-structures format
\r
1922 // Default constructor; does no initialization
\r
1924 __forceinline Matrix4( ) { };
\r
1926 // Copy a 4x4 matrix
\r
1928 __forceinline Matrix4( const Matrix4 & mat );
\r
1930 // Construct a 4x4 matrix containing the specified columns
\r
1932 __forceinline Matrix4( const Vector4 &col0, const Vector4 &col1, const Vector4 &col2, const Vector4 &col3 );
\r
1934 // Construct a 4x4 matrix from a 3x4 transformation matrix
\r
1936 explicit __forceinline Matrix4( const Transform3 & mat );
\r
1938 // Construct a 4x4 matrix from a 3x3 matrix and a 3-D vector
\r
1940 __forceinline Matrix4( const Matrix3 & mat, const Vector3 &translateVec );
\r
1942 // Construct a 4x4 matrix from a unit-length quaternion and a 3-D vector
\r
1944 __forceinline Matrix4( const Quat &unitQuat, const Vector3 &translateVec );
\r
1946 // Set all elements of a 4x4 matrix to the same scalar value
\r
1948 explicit __forceinline Matrix4( float scalar );
\r
1950 // Set all elements of a 4x4 matrix to the same scalar value (scalar data contained in vector data type)
\r
1952 explicit __forceinline Matrix4( const floatInVec &scalar );
\r
1954 // Assign one 4x4 matrix to another
\r
1956 __forceinline Matrix4 & operator =( const Matrix4 & mat );
\r
1958 // Set the upper-left 3x3 submatrix
\r
1960 // This function does not change the bottom row elements.
\r
1962 __forceinline Matrix4 & setUpper3x3( const Matrix3 & mat3 );
\r
1964 // Get the upper-left 3x3 submatrix of a 4x4 matrix
\r
1966 __forceinline const Matrix3 getUpper3x3( ) const;
\r
1968 // Set translation component
\r
1970 // This function does not change the bottom row elements.
\r
1972 __forceinline Matrix4 & setTranslation( const Vector3 &translateVec );
\r
1974 // Get the translation component of a 4x4 matrix
\r
1976 __forceinline const Vector3 getTranslation( ) const;
\r
1978 // Set column 0 of a 4x4 matrix
\r
1980 __forceinline Matrix4 & setCol0( const Vector4 &col0 );
\r
1982 // Set column 1 of a 4x4 matrix
\r
1984 __forceinline Matrix4 & setCol1( const Vector4 &col1 );
\r
1986 // Set column 2 of a 4x4 matrix
\r
1988 __forceinline Matrix4 & setCol2( const Vector4 &col2 );
\r
1990 // Set column 3 of a 4x4 matrix
\r
1992 __forceinline Matrix4 & setCol3( const Vector4 &col3 );
\r
1994 // Get column 0 of a 4x4 matrix
\r
1996 __forceinline const Vector4 getCol0( ) const;
\r
1998 // Get column 1 of a 4x4 matrix
\r
2000 __forceinline const Vector4 getCol1( ) const;
\r
2002 // Get column 2 of a 4x4 matrix
\r
2004 __forceinline const Vector4 getCol2( ) const;
\r
2006 // Get column 3 of a 4x4 matrix
\r
2008 __forceinline const Vector4 getCol3( ) const;
\r
2010 // Set the column of a 4x4 matrix referred to by the specified index
\r
2012 __forceinline Matrix4 & setCol( int col, const Vector4 &vec );
\r
2014 // Set the row of a 4x4 matrix referred to by the specified index
\r
2016 __forceinline Matrix4 & setRow( int row, const Vector4 &vec );
\r
2018 // Get the column of a 4x4 matrix referred to by the specified index
\r
2020 __forceinline const Vector4 getCol( int col ) const;
\r
2022 // Get the row of a 4x4 matrix referred to by the specified index
\r
2024 __forceinline const Vector4 getRow( int row ) const;
\r
2026 // Subscripting operator to set or get a column
\r
2028 __forceinline Vector4 & operator []( int col );
\r
2030 // Subscripting operator to get a column
\r
2032 __forceinline const Vector4 operator []( int col ) const;
\r
2034 // Set the element of a 4x4 matrix referred to by column and row indices
\r
2036 __forceinline Matrix4 & setElem( int col, int row, float val );
\r
2038 // Set the element of a 4x4 matrix referred to by column and row indices (scalar data contained in vector data type)
\r
2040 __forceinline Matrix4 & setElem( int col, int row, const floatInVec &val );
\r
2042 // Get the element of a 4x4 matrix referred to by column and row indices
\r
2044 __forceinline const floatInVec getElem( int col, int row ) const;
\r
2046 // Add two 4x4 matrices
\r
2048 __forceinline const Matrix4 operator +( const Matrix4 & mat ) const;
\r
2050 // Subtract a 4x4 matrix from another 4x4 matrix
\r
2052 __forceinline const Matrix4 operator -( const Matrix4 & mat ) const;
\r
2054 // Negate all elements of a 4x4 matrix
\r
2056 __forceinline const Matrix4 operator -( ) const;
\r
2058 // Multiply a 4x4 matrix by a scalar
\r
2060 __forceinline const Matrix4 operator *( float scalar ) const;
\r
2062 // Multiply a 4x4 matrix by a scalar (scalar data contained in vector data type)
\r
2064 __forceinline const Matrix4 operator *( const floatInVec &scalar ) const;
\r
2066 // Multiply a 4x4 matrix by a 4-D vector
\r
2068 __forceinline const Vector4 operator *( const Vector4 &vec ) const;
\r
2070 // Multiply a 4x4 matrix by a 3-D vector
\r
2072 __forceinline const Vector4 operator *( const Vector3 &vec ) const;
\r
2074 // Multiply a 4x4 matrix by a 3-D point
\r
2076 __forceinline const Vector4 operator *( const Point3 &pnt ) const;
\r
2078 // Multiply two 4x4 matrices
\r
2080 __forceinline const Matrix4 operator *( const Matrix4 & mat ) const;
\r
2082 // Multiply a 4x4 matrix by a 3x4 transformation matrix
\r
2084 __forceinline const Matrix4 operator *( const Transform3 & tfrm ) const;
\r
2086 // Perform compound assignment and addition with a 4x4 matrix
\r
2088 __forceinline Matrix4 & operator +=( const Matrix4 & mat );
\r
2090 // Perform compound assignment and subtraction by a 4x4 matrix
\r
2092 __forceinline Matrix4 & operator -=( const Matrix4 & mat );
\r
2094 // Perform compound assignment and multiplication by a scalar
\r
2096 __forceinline Matrix4 & operator *=( float scalar );
\r
2098 // Perform compound assignment and multiplication by a scalar (scalar data contained in vector data type)
\r
2100 __forceinline Matrix4 & operator *=( const floatInVec &scalar );
\r
2102 // Perform compound assignment and multiplication by a 4x4 matrix
\r
2104 __forceinline Matrix4 & operator *=( const Matrix4 & mat );
\r
2106 // Perform compound assignment and multiplication by a 3x4 transformation matrix
\r
2108 __forceinline Matrix4 & operator *=( const Transform3 & tfrm );
\r
2110 // Construct an identity 4x4 matrix
\r
2112 static __forceinline const Matrix4 identity( );
\r
2114 // Construct a 4x4 matrix to rotate around the x axis
\r
2116 static __forceinline const Matrix4 rotationX( float radians );
\r
2118 // Construct a 4x4 matrix to rotate around the y axis
\r
2120 static __forceinline const Matrix4 rotationY( float radians );
\r
2122 // Construct a 4x4 matrix to rotate around the z axis
\r
2124 static __forceinline const Matrix4 rotationZ( float radians );
\r
2126 // Construct a 4x4 matrix to rotate around the x axis (scalar data contained in vector data type)
\r
2128 static __forceinline const Matrix4 rotationX( const floatInVec &radians );
\r
2130 // Construct a 4x4 matrix to rotate around the y axis (scalar data contained in vector data type)
\r
2132 static __forceinline const Matrix4 rotationY( const floatInVec &radians );
\r
2134 // Construct a 4x4 matrix to rotate around the z axis (scalar data contained in vector data type)
\r
2136 static __forceinline const Matrix4 rotationZ( const floatInVec &radians );
\r
2138 // Construct a 4x4 matrix to rotate around the x, y, and z axes
\r
2140 static __forceinline const Matrix4 rotationZYX( const Vector3 &radiansXYZ );
\r
2142 // Construct a 4x4 matrix to rotate around a unit-length 3-D vector
\r
2144 static __forceinline const Matrix4 rotation( float radians, const Vector3 &unitVec );
\r
2146 // Construct a 4x4 matrix to rotate around a unit-length 3-D vector (scalar data contained in vector data type)
\r
2148 static __forceinline const Matrix4 rotation( const floatInVec &radians, const Vector3 &unitVec );
\r
2150 // Construct a rotation matrix from a unit-length quaternion
\r
2152 static __forceinline const Matrix4 rotation( const Quat &unitQuat );
\r
2154 // Construct a 4x4 matrix to perform scaling
\r
2156 static __forceinline const Matrix4 scale( const Vector3 &scaleVec );
\r
2158 // Construct a 4x4 matrix to perform translation
\r
2160 static __forceinline const Matrix4 translation( const Vector3 &translateVec );
\r
2162 // Construct viewing matrix based on eye, position looked at, and up direction
\r
2164 static __forceinline const Matrix4 lookAt( const Point3 &eyePos, const Point3 &lookAtPos, const Vector3 &upVec );
\r
2166 // Construct a perspective projection matrix
\r
2168 static __forceinline const Matrix4 perspective( float fovyRadians, float aspect, float zNear, float zFar );
\r
2170 // Construct a perspective projection matrix based on frustum
\r
2172 static __forceinline const Matrix4 frustum( float left, float right, float bottom, float top, float zNear, float zFar );
\r
2174 // Construct an orthographic projection matrix
\r
2176 static __forceinline const Matrix4 orthographic( float left, float right, float bottom, float top, float zNear, float zFar );
\r
2179 // Multiply a 4x4 matrix by a scalar
\r
2181 __forceinline const Matrix4 operator *( float scalar, const Matrix4 & mat );
\r
2183 // Multiply a 4x4 matrix by a scalar (scalar data contained in vector data type)
\r
2185 __forceinline const Matrix4 operator *( const floatInVec &scalar, const Matrix4 & mat );
\r
2187 // Append (post-multiply) a scale transformation to a 4x4 matrix
\r
2189 // Faster than creating and multiplying a scale transformation matrix.
\r
2191 __forceinline const Matrix4 appendScale( const Matrix4 & mat, const Vector3 &scaleVec );
\r
2193 // Prepend (pre-multiply) a scale transformation to a 4x4 matrix
\r
2195 // Faster than creating and multiplying a scale transformation matrix.
\r
2197 __forceinline const Matrix4 prependScale( const Vector3 &scaleVec, const Matrix4 & mat );
\r
2199 // Multiply two 4x4 matrices per element
\r
2201 __forceinline const Matrix4 mulPerElem( const Matrix4 & mat0, const Matrix4 & mat1 );
\r
2203 // Compute the absolute value of a 4x4 matrix per element
\r
2205 __forceinline const Matrix4 absPerElem( const Matrix4 & mat );
\r
2207 // Transpose of a 4x4 matrix
\r
2209 __forceinline const Matrix4 transpose( const Matrix4 & mat );
\r
2211 // Compute the inverse of a 4x4 matrix
\r
2213 // Result is unpredictable when the determinant of mat is equal to or near 0.
\r
2215 __forceinline const Matrix4 inverse( const Matrix4 & mat );
\r
2217 // Compute the inverse of a 4x4 matrix, which is expected to be an affine matrix
\r
2219 // This can be used to achieve better performance than a general inverse when the specified 4x4 matrix meets the given restrictions. The result is unpredictable when the determinant of mat is equal to or near 0.
\r
2221 __forceinline const Matrix4 affineInverse( const Matrix4 & mat );
\r
2223 // Compute the inverse of a 4x4 matrix, which is expected to be an affine matrix with an orthogonal upper-left 3x3 submatrix
\r
2225 // This can be used to achieve better performance than a general inverse when the specified 4x4 matrix meets the given restrictions.
\r
2227 __forceinline const Matrix4 orthoInverse( const Matrix4 & mat );
\r
2229 // Determinant of a 4x4 matrix
\r
2231 __forceinline const floatInVec determinant( const Matrix4 & mat );
\r
2233 // Conditionally select between two 4x4 matrices
\r
2235 // This function uses a conditional select instruction to avoid a branch.
\r
2236 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
2237 // Use the boolInVec version for better performance.
\r
2239 __forceinline const Matrix4 select( const Matrix4 & mat0, const Matrix4 & mat1, bool select1 );
\r
2241 // Conditionally select between two 4x4 matrices (scalar data contained in vector data type)
\r
2243 // This function uses a conditional select instruction to avoid a branch.
\r
2245 __forceinline const Matrix4 select( const Matrix4 & mat0, const Matrix4 & mat1, const boolInVec &select1 );
\r
2247 #ifdef _VECTORMATH_DEBUG
\r
2249 // Print a 4x4 matrix
\r
2251 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
2253 __forceinline void print( const Matrix4 & mat );
\r
2255 // Print a 4x4 matrix and an associated string identifier
\r
2257 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
2259 __forceinline void print( const Matrix4 & mat, const char * name );
\r
2263 // A 3x4 transformation matrix in array-of-structures format
\r
2273 // Default constructor; does no initialization
\r
2275 __forceinline Transform3( ) { };
\r
2277 // Copy a 3x4 transformation matrix
\r
2279 __forceinline Transform3( const Transform3 & tfrm );
\r
2281 // Construct a 3x4 transformation matrix containing the specified columns
\r
2283 __forceinline Transform3( const Vector3 &col0, const Vector3 &col1, const Vector3 &col2, const Vector3 &col3 );
\r
2285 // Construct a 3x4 transformation matrix from a 3x3 matrix and a 3-D vector
\r
2287 __forceinline Transform3( const Matrix3 & tfrm, const Vector3 &translateVec );
\r
2289 // Construct a 3x4 transformation matrix from a unit-length quaternion and a 3-D vector
\r
2291 __forceinline Transform3( const Quat &unitQuat, const Vector3 &translateVec );
\r
2293 // Set all elements of a 3x4 transformation matrix to the same scalar value
\r
2295 explicit __forceinline Transform3( float scalar );
\r
2297 // Set all elements of a 3x4 transformation matrix to the same scalar value (scalar data contained in vector data type)
\r
2299 explicit __forceinline Transform3( const floatInVec &scalar );
\r
2301 // Assign one 3x4 transformation matrix to another
\r
2303 __forceinline Transform3 & operator =( const Transform3 & tfrm );
\r
2305 // Set the upper-left 3x3 submatrix
\r
2307 __forceinline Transform3 & setUpper3x3( const Matrix3 & mat3 );
\r
2309 // Get the upper-left 3x3 submatrix of a 3x4 transformation matrix
\r
2311 __forceinline const Matrix3 getUpper3x3( ) const;
\r
2313 // Set translation component
\r
2315 __forceinline Transform3 & setTranslation( const Vector3 &translateVec );
\r
2317 // Get the translation component of a 3x4 transformation matrix
\r
2319 __forceinline const Vector3 getTranslation( ) const;
\r
2321 // Set column 0 of a 3x4 transformation matrix
\r
2323 __forceinline Transform3 & setCol0( const Vector3 &col0 );
\r
2325 // Set column 1 of a 3x4 transformation matrix
\r
2327 __forceinline Transform3 & setCol1( const Vector3 &col1 );
\r
2329 // Set column 2 of a 3x4 transformation matrix
\r
2331 __forceinline Transform3 & setCol2( const Vector3 &col2 );
\r
2333 // Set column 3 of a 3x4 transformation matrix
\r
2335 __forceinline Transform3 & setCol3( const Vector3 &col3 );
\r
2337 // Get column 0 of a 3x4 transformation matrix
\r
2339 __forceinline const Vector3 getCol0( ) const;
\r
2341 // Get column 1 of a 3x4 transformation matrix
\r
2343 __forceinline const Vector3 getCol1( ) const;
\r
2345 // Get column 2 of a 3x4 transformation matrix
\r
2347 __forceinline const Vector3 getCol2( ) const;
\r
2349 // Get column 3 of a 3x4 transformation matrix
\r
2351 __forceinline const Vector3 getCol3( ) const;
\r
2353 // Set the column of a 3x4 transformation matrix referred to by the specified index
\r
2355 __forceinline Transform3 & setCol( int col, const Vector3 &vec );
\r
2357 // Set the row of a 3x4 transformation matrix referred to by the specified index
\r
2359 __forceinline Transform3 & setRow( int row, const Vector4 &vec );
\r
2361 // Get the column of a 3x4 transformation matrix referred to by the specified index
\r
2363 __forceinline const Vector3 getCol( int col ) const;
\r
2365 // Get the row of a 3x4 transformation matrix referred to by the specified index
\r
2367 __forceinline const Vector4 getRow( int row ) const;
\r
2369 // Subscripting operator to set or get a column
\r
2371 __forceinline Vector3 & operator []( int col );
\r
2373 // Subscripting operator to get a column
\r
2375 __forceinline const Vector3 operator []( int col ) const;
\r
2377 // Set the element of a 3x4 transformation matrix referred to by column and row indices
\r
2379 __forceinline Transform3 & setElem( int col, int row, float val );
\r
2381 // Set the element of a 3x4 transformation matrix referred to by column and row indices (scalar data contained in vector data type)
\r
2383 __forceinline Transform3 & setElem( int col, int row, const floatInVec &val );
\r
2385 // Get the element of a 3x4 transformation matrix referred to by column and row indices
\r
2387 __forceinline const floatInVec getElem( int col, int row ) const;
\r
2389 // Multiply a 3x4 transformation matrix by a 3-D vector
\r
2391 __forceinline const Vector3 operator *( const Vector3 &vec ) const;
\r
2393 // Multiply a 3x4 transformation matrix by a 3-D point
\r
2395 __forceinline const Point3 operator *( const Point3 &pnt ) const;
\r
2397 // Multiply two 3x4 transformation matrices
\r
2399 __forceinline const Transform3 operator *( const Transform3 & tfrm ) const;
\r
2401 // Perform compound assignment and multiplication by a 3x4 transformation matrix
\r
2403 __forceinline Transform3 & operator *=( const Transform3 & tfrm );
\r
2405 // Construct an identity 3x4 transformation matrix
\r
2407 static __forceinline const Transform3 identity( );
\r
2409 // Construct a 3x4 transformation matrix to rotate around the x axis
\r
2411 static __forceinline const Transform3 rotationX( float radians );
\r
2413 // Construct a 3x4 transformation matrix to rotate around the y axis
\r
2415 static __forceinline const Transform3 rotationY( float radians );
\r
2417 // Construct a 3x4 transformation matrix to rotate around the z axis
\r
2419 static __forceinline const Transform3 rotationZ( float radians );
\r
2421 // Construct a 3x4 transformation matrix to rotate around the x axis (scalar data contained in vector data type)
\r
2423 static __forceinline const Transform3 rotationX( const floatInVec &radians );
\r
2425 // Construct a 3x4 transformation matrix to rotate around the y axis (scalar data contained in vector data type)
\r
2427 static __forceinline const Transform3 rotationY( const floatInVec &radians );
\r
2429 // Construct a 3x4 transformation matrix to rotate around the z axis (scalar data contained in vector data type)
\r
2431 static __forceinline const Transform3 rotationZ( const floatInVec &radians );
\r
2433 // Construct a 3x4 transformation matrix to rotate around the x, y, and z axes
\r
2435 static __forceinline const Transform3 rotationZYX( const Vector3 &radiansXYZ );
\r
2437 // Construct a 3x4 transformation matrix to rotate around a unit-length 3-D vector
\r
2439 static __forceinline const Transform3 rotation( float radians, const Vector3 &unitVec );
\r
2441 // Construct a 3x4 transformation matrix to rotate around a unit-length 3-D vector (scalar data contained in vector data type)
\r
2443 static __forceinline const Transform3 rotation( const floatInVec &radians, const Vector3 &unitVec );
\r
2445 // Construct a rotation matrix from a unit-length quaternion
\r
2447 static __forceinline const Transform3 rotation( const Quat &unitQuat );
\r
2449 // Construct a 3x4 transformation matrix to perform scaling
\r
2451 static __forceinline const Transform3 scale( const Vector3 &scaleVec );
\r
2453 // Construct a 3x4 transformation matrix to perform translation
\r
2455 static __forceinline const Transform3 translation( const Vector3 &translateVec );
\r
2458 // Append (post-multiply) a scale transformation to a 3x4 transformation matrix
\r
2460 // Faster than creating and multiplying a scale transformation matrix.
\r
2462 __forceinline const Transform3 appendScale( const Transform3 & tfrm, const Vector3 &scaleVec );
\r
2464 // Prepend (pre-multiply) a scale transformation to a 3x4 transformation matrix
\r
2466 // Faster than creating and multiplying a scale transformation matrix.
\r
2468 __forceinline const Transform3 prependScale( const Vector3 &scaleVec, const Transform3 & tfrm );
\r
2470 // Multiply two 3x4 transformation matrices per element
\r
2472 __forceinline const Transform3 mulPerElem( const Transform3 & tfrm0, const Transform3 & tfrm1 );
\r
2474 // Compute the absolute value of a 3x4 transformation matrix per element
\r
2476 __forceinline const Transform3 absPerElem( const Transform3 & tfrm );
\r
2478 // Inverse of a 3x4 transformation matrix
\r
2480 // Result is unpredictable when the determinant of the left 3x3 submatrix is equal to or near 0.
\r
2482 __forceinline const Transform3 inverse( const Transform3 & tfrm );
\r
2484 // Compute the inverse of a 3x4 transformation matrix, expected to have an orthogonal upper-left 3x3 submatrix
\r
2486 // This can be used to achieve better performance than a general inverse when the specified 3x4 transformation matrix meets the given restrictions.
\r
2488 __forceinline const Transform3 orthoInverse( const Transform3 & tfrm );
\r
2490 // Conditionally select between two 3x4 transformation matrices
\r
2492 // This function uses a conditional select instruction to avoid a branch.
\r
2493 // However, the transfer of select1 to a VMX register may use more processing time than a branch.
\r
2494 // Use the boolInVec version for better performance.
\r
2496 __forceinline const Transform3 select( const Transform3 & tfrm0, const Transform3 & tfrm1, bool select1 );
\r
2498 // Conditionally select between two 3x4 transformation matrices (scalar data contained in vector data type)
\r
2500 // This function uses a conditional select instruction to avoid a branch.
\r
2502 __forceinline const Transform3 select( const Transform3 & tfrm0, const Transform3 & tfrm1, const boolInVec &select1 );
\r
2504 #ifdef _VECTORMATH_DEBUG
\r
2506 // Print a 3x4 transformation matrix
\r
2508 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
2510 __forceinline void print( const Transform3 & tfrm );
\r
2512 // Print a 3x4 transformation matrix and an associated string identifier
\r
2514 // Function is only defined when _VECTORMATH_DEBUG is defined.
\r
2516 __forceinline void print( const Transform3 & tfrm, const char * name );
\r
2520 } // namespace Aos
\r
2521 } // namespace Vectormath
\r
2523 #include "vec_aos.h"
\r
2524 #include "quat_aos.h"
\r
2525 #include "mat_aos.h"
\r