/************************************************************************* * * * Open Dynamics Engine, Copyright (C) 2001,2002 Russell L. Smith. * * All rights reserved. Email: russ@q12.org Web: www.q12.org * * * * This library is free software; you can redistribute it and/or * * modify it under the terms of EITHER: * * (1) The GNU Lesser General Public License as published by the Free * * Software Foundation; either version 2.1 of the License, or (at * * your option) any later version. The text of the GNU Lesser * * General Public License is included with this library in the * * file LICENSE.TXT. * * (2) The BSD-style license that is included with this library in * * the file LICENSE-BSD.TXT. * * * * This library is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the files * * LICENSE.TXT and LICENSE-BSD.TXT for more details. * * * *************************************************************************/ /* * optimized and unoptimized vector and matrix functions * (inlined private versions) */ #ifndef _ODE__PRIVATE_MATRIX_H_ #define _ODE__PRIVATE_MATRIX_H_ #include #ifdef __cplusplus template ODE_INLINE void dxtSetZero (element_type *a, sizeint n) { element_type *const aend = a + n * a_stride; for (element_type *acurr = a; acurr != aend; acurr += a_stride) { *acurr = (element_type)0; } } template ODE_INLINE void dxSetZero (element_type *a, sizeint n) { dxtSetZero<1>(a, n); } template ODE_INLINE void dxSetValue (element_type *a, sizeint n, element_type value) { element_type *const aend = a + n; for (element_type *acurr = a; acurr != aend; ++acurr) { *acurr = value; } } #else // #ifndef __cplusplus ODE_PURE_INLINE void dxSetZero (dReal *a, sizeint n) { dReal *const aend = a + n; dReal *acurr; for (acurr = a; acurr != aend; ++acurr) { *acurr = 0; } } ODE_PURE_INLINE void dxSetValue (dReal *a, sizeint n, dReal value) { dReal *const aend = a + n; dReal *acurr; for (acurr = a; acurr != aend; ++acurr) { *acurr = value; } } #endif // #ifdef __cplusplus dReal dxDot (const dReal *a, const dReal *b, unsigned n); void dxMultiply0 (dReal *A, const dReal *B, const dReal *C, unsigned p, unsigned q, unsigned r); void dxMultiply1 (dReal *A, const dReal *B, const dReal *C, unsigned p, unsigned q, unsigned r); void dxMultiply2 (dReal *A, const dReal *B, const dReal *C, unsigned p, unsigned q, unsigned r); int dxFactorCholesky (dReal *A, unsigned n, void *tmpbuf); void dxSolveCholesky (const dReal *L, dReal *b, unsigned n, void *tmpbuf); int dxInvertPDMatrix (const dReal *A, dReal *Ainv, unsigned n, void *tmpbuf); int dxIsPositiveDefinite (const dReal *A, unsigned n, void *tmpbuf); void dxLDLTAddTL (dReal *L, dReal *d, const dReal *a, unsigned n, unsigned nskip, void *tmpbuf); void dxLDLTRemove (dReal **A, const unsigned *p, dReal *L, dReal *d, unsigned n1, unsigned n2, unsigned r, unsigned nskip, void *tmpbuf); void dxRemoveRowCol (dReal *A, unsigned n, unsigned nskip, unsigned r); ODE_PURE_INLINE sizeint dxEstimateFactorCholeskyTmpbufSize(unsigned n) { return dPAD(n) * sizeof(dReal); } ODE_PURE_INLINE sizeint dxEstimateSolveCholeskyTmpbufSize(unsigned n) { return dPAD(n) * sizeof(dReal); } ODE_PURE_INLINE sizeint dxEstimateInvertPDMatrixTmpbufSize(unsigned n) { sizeint FactorCholesky_size = dxEstimateFactorCholeskyTmpbufSize(n); sizeint SolveCholesky_size = dxEstimateSolveCholeskyTmpbufSize(n); sizeint MaxCholesky_size = FactorCholesky_size > SolveCholesky_size ? FactorCholesky_size : SolveCholesky_size; return (sizeint)dPAD(n) * (n + 1) * sizeof(dReal) + MaxCholesky_size; } ODE_PURE_INLINE sizeint dxEstimateIsPositiveDefiniteTmpbufSize(unsigned n) { return (sizeint)dPAD(n) * n * sizeof(dReal) + dxEstimateFactorCholeskyTmpbufSize(n); } ODE_PURE_INLINE sizeint dxEstimateLDLTAddTLTmpbufSize(unsigned nskip) { return nskip * (2 * sizeof(dReal)); } ODE_PURE_INLINE sizeint dxEstimateLDLTRemoveTmpbufSize(unsigned n2, unsigned nskip) { return n2 * sizeof(dReal) + dxEstimateLDLTAddTLTmpbufSize(nskip); } /* For internal use */ #define dSetZero(a, n) dxSetZero(a, n) #define dSetValue(a, n, value) dxSetValue(a, n, value) #define dDot(a, b, n) dxDot(a, b, n) #define dMultiply0(A, B, C, p, q, r) dxMultiply0(A, B, C, p, q, r) #define dMultiply1(A, B, C, p, q, r) dxMultiply1(A, B, C, p, q, r) #define dMultiply2(A, B, C, p, q, r) dxMultiply2(A, B, C, p, q, r) #define dFactorCholesky(A, n, tmpbuf) dxFactorCholesky(A, n, tmpbuf) #define dSolveCholesky(L, b, n, tmpbuf) dxSolveCholesky(L, b, n, tmpbuf) #define dInvertPDMatrix(A, Ainv, n, tmpbuf) dxInvertPDMatrix(A, Ainv, n, tmpbuf) #define dIsPositiveDefinite(A, n, tmpbuf) dxIsPositiveDefinite(A, n, tmpbuf) #define dLDLTAddTL(L, d, a, n, nskip, tmpbuf) dxLDLTAddTL(L, d, a, n, nskip, tmpbuf) #define dLDLTRemove(A, p, L, d, n1, n2, r, nskip, tmpbuf) dxLDLTRemove(A, p, L, d, n1, n2, r, nskip, tmpbuf) #define dRemoveRowCol(A, n, nskip, r) dxRemoveRowCol(A, n, nskip, r) #define dEstimateFactorCholeskyTmpbufSize(n) dxEstimateFactorCholeskyTmpbufSize(n) #define dEstimateSolveCholeskyTmpbufSize(n) dxEstimateSolveCholeskyTmpbufSize(n) #define dEstimateInvertPDMatrixTmpbufSize(n) dxEstimateInvertPDMatrixTmpbufSize(n) #define dEstimateIsPositiveDefiniteTmpbufSize(n) dxEstimateIsPositiveDefiniteTmpbufSize(n) #define dEstimateLDLTAddTLTmpbufSize(nskip) dxEstimateLDLTAddTLTmpbufSize(nskip) #define dEstimateLDLTRemoveTmpbufSize(n2, nskip) dxEstimateLDLTRemoveTmpbufSize(n2, nskip) #endif // #ifndef _ODE__PRIVATE_MATRIX_H_