TREX/qmckl
1
0
Fork 0
mirror of https://github.com/TREX-CoE/qmckl.git synced 2024-12-23 04:44:03 +01:00
Code Issues Projects Releases Wiki Activity

Optimized polynomials

Browse Source
This commit is contained in:
Anthony Scemama 2022-02-16 19:40:14 +01:00
parent e90e9a531c
commit 733d941c30
1 changed files with 169 additions and 144 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

309
org/qmckl_ao.org
View File

@ -4530,153 +4530,176 @@ qmckl_ao_polynomial_transp_vgl_hpc (const qmckl_context context,
int32_t m; int32_t m;
double* const vgl1 = VGL;
double* const vgl2 = VGL + ldv;
double* const vgl3 = VGL + (ldv << 1);
double* const vgl4 = VGL + 3*ldv;
double* const vgl5 = VGL + (ldv << 2);
switch (lmax) { switch (lmax) {
case 0: case 0:
{ {
if (ldv < 1) return QMCKL_INVALID_ARG_9; if (ldv < 1) return QMCKL_INVALID_ARG_9;
L[0] = 0; L[0] = 0; L[1] = 0; L[2] = 0;
L[ldl] = 0;
L[ldl << 1] = 0;
m=1;
vgl1[0] = 1.0; VGL[0 ] = 1.0;
vgl2[0] = 0.0; VGL[ldv ] = 0.0;
vgl3[0] = 0.0; VGL[ldv<<1 ] = 0.0;
vgl4[0] = 0.0; VGL[ldv<<2-1] = 0.0;
vgl5[0] = 0.0; VGL[ldv<<2 ] = 0.0;
m=1;
break; break;
} }
case 1: case 1:
{ {
if (ldv < 4) return QMCKL_INVALID_ARG_9; if (ldv < 4) return QMCKL_INVALID_ARG_9;
const double Y[3] = { X[0]-R[0],
X[1]-R[1],
X[2]-R[2] };
int32_t* const l[4] = {L, L+ldl, L+(ldl<<1), L+3*ldl}; if (ldl == 3) {
for (int32_t i=0 ; i<4 ; ++i) {
l[i][0] = 0;
l[i][1] = 0;
l[i][2] = 0;
}
l[1][0] = 1;
l[2][1] = 1;
l[3][2] = 1;
for (int32_t k=0 ; k<4 ; ++k) { const int32_t ltmp[12] = {0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1};
vgl2[k] = 0.0; for (int i=0 ; i<12 ; ++i)
vgl3[k] = 0.0; L[i] = ltmp[i];
vgl4[k] = 0.0;
vgl5[k] = 0.0; } else {
int32_t* const l[4] = {L, L+ldl, L+(ldl<<1), L+3*ldl};
l[0][0] = 0; l[0][1] = 0; l[0][2] = 0;
l[1][0] = 1; l[1][1] = 0; l[1][2] = 0;
l[2][0] = 0; l[2][1] = 1; l[2][2] = 0;
l[3][0] = 0; l[3][1] = 0; l[3][2] = 1;
}
if (ldv == 4) {
const double tmp[20] = {1.0, X[0]-R[0], X[1]-R[1], X[2]-R[2], 0.0,
1.0, 0.0, 0.0, 0.0, 0.0,
1.0, 0.0, 0.0, 0.0, 0.0,
1.0, 0.0, 0.0, 0.0, 0.0};
for (int i=0 ; i<20 ; ++i)
VGL[i] = tmp[i];
} else {
double* const vgl1 = VGL;
double* const vgl2 = VGL + ldv;
double* const vgl3 = VGL + (ldv << 1);
double* const vgl4 = VGL + 3*ldv;
double* const vgl5 = VGL + (ldv << 2);
for (int32_t k=0 ; k<4 ; ++k) {
vgl2[k] = 0.0;
vgl3[k] = 0.0;
vgl4[k] = 0.0;
vgl5[k] = 0.0;
}
vgl1[0] = 1.0;
vgl1[1] = X[0]-R[0];
vgl1[2] = X[1]-R[1];
vgl1[3] = X[2]-R[2];
vgl2[1] = 1.0;
vgl3[2] = 1.0;
vgl4[3] = 1.0;
} }
vgl1[0] = 1.0;
vgl1[1] = Y[0];
vgl1[2] = Y[1];
vgl1[3] = Y[2];
vgl2[1] = 1.0;
vgl3[2] = 1.0;
vgl4[3] = 1.0;
m=4; m=4;
break; break;
} }
case 2: case 2:
{ {
if (ldv < 10) return QMCKL_INVALID_ARG_9; if (ldv < 10) return QMCKL_INVALID_ARG_9;
if (ldl == 3) {
const int32_t ltmp[30] = {0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1,
2, 0, 0, 1, 1, 0, 1, 0, 1, 0, 2, 0,
0, 1, 1, 0, 0, 2};
for (int i=0 ; i<30 ; ++i)
L[i] = ltmp[i];
} else {
int32_t* l[10];
for (int32_t i=0 ; i<10 ; ++i) {
l[i] = L + i*ldl;
}
l[0][0] = 0; l[0][1] = 0; l[0][2] = 0;
l[1][0] = 1; l[1][1] = 0; l[1][2] = 0;
l[2][0] = 0; l[2][1] = 1; l[2][2] = 0;
l[3][0] = 0; l[3][1] = 0; l[3][2] = 1;
l[4][0] = 2; l[4][1] = 0; l[4][2] = 0;
l[5][0] = 1; l[5][1] = 1; l[5][2] = 0;
l[6][0] = 1; l[6][1] = 0; l[6][2] = 1;
l[7][0] = 0; l[7][1] = 2; l[7][2] = 0;
l[8][0] = 0; l[8][1] = 1; l[8][2] = 1;
l[9][0] = 0; l[9][1] = 0; l[9][2] = 2;
}
const double Y[3] = { X[0]-R[0], const double Y[3] = { X[0]-R[0],
X[1]-R[1], X[1]-R[1],
X[2]-R[2] }; X[2]-R[2] };
double pows[3][5] = { {1.0, 1.0, 1.0, Y[0], Y[0]*Y[0]}, if (ldv == 50) {
{1.0, 1.0, 1.0, Y[1], Y[1]*Y[1]}, const double tmp[50] = {
{1.0, 1.0, 1.0, Y[2], Y[2]*Y[2]} }; 1.0, Y[0], Y[1], Y[2], Y[0] * Y[0],
Y[0] * Y[1], Y[0] * Y[2], Y[1] * Y[1], Y[1] * Y[2], Y[2] * Y[2],
0.0, 1.0, 0.0, 0.0, Y[0] + Y[0],
Y[1], Y[2], 0.0, 0.0, 0.0,
0.0, 0.0, 1.0, 0.0, 0.0,
Y[0], 0.0, Y[1] + Y[1], Y[2], 0.0,
0.0, 0.0, 0.0, 1.0, 0.0,
0.0, Y[0], 0.0, Y[1], Y[2] + Y[2],
0.0, 0.0, 0.0, 0.0, 2.0,
0.0, 0.0, 2.0, 0., 2.0 };
for (int i=0 ; i<50 ; ++i)
VGL[i] = tmp[i];
} else {
double* const vgl1 = VGL;
double* const vgl2 = VGL + ldv;
double* const vgl3 = VGL + (ldv << 1);
double* const vgl4 = VGL + 3*ldv;
double* const vgl5 = VGL + (ldv << 2);
int32_t* l[10]; vgl1[0] = 1.0 ; vgl1[1] = Y[0] ; vgl1[2] = Y[1];
for (int32_t i=0 ; i<10 ; ++i) { vgl1[3] = Y[2] ; vgl1[4] = Y[0]*Y[0]; vgl1[5] = Y[0]*Y[1];
l[i] = L + i*ldl; vgl1[6] = Y[0]*Y[2]; vgl1[7] = Y[1]*Y[1]; vgl1[8] = Y[1]*Y[2];
} vgl1[9] = Y[2]*Y[2];
for (int32_t i=0 ; i<4 ; ++i) { vgl2[0] = 0.0 ; vgl2[1] = 1.0 ; vgl2[2] = 0.0 ;
l[i][0] = 0; vgl2[3] = 0.0 ; vgl2[4] = Y[0]+Y[0]; vgl2[5] = Y[1];
l[i][1] = 0; vgl2[6] = Y[2]; vgl2[7] = 0.0 ; vgl2[8] = 0.0 ;
l[i][2] = 0; vgl2[9] = 0.0 ;
}
l[1][0] = 1; vgl3[0] = 0.0; vgl3[1] = 0.0 ; vgl3[2] = 1.0 ;
l[2][1] = 1; vgl3[3] = 0.0; vgl3[4] = 0.0 ; vgl3[5] = Y[0];
l[3][2] = 1; vgl3[6] = 0.0; vgl3[7] = Y[1]+Y[1]; vgl3[8] = Y[2];
vgl3[9] = 0.0;
for (int32_t k=0 ; k<4 ; ++k) {
vgl2[k] = 0.0; vgl4[0] = 0.0 ; vgl4[1] = 0.0; vgl4[2] = 0.0 ;
vgl3[k] = 0.0; vgl4[3] = 1.0 ; vgl4[4] = 0.0; vgl4[5] = 0.0 ;
vgl4[k] = 0.0; vgl4[6] = Y[0] ; vgl4[7] = 0.0; vgl4[8] = Y[1];
vgl5[k] = 0.0; vgl4[9] = Y[2]+Y[2];
}
vgl1[0] = 1.0; vgl5[0] = 0.0; vgl5[1] = 0.0; vgl5[2] = 0.0;
vgl1[1] = Y[0]; vgl5[3] = 0.0; vgl5[4] = 2.0; vgl5[5] = 0.0;
vgl1[2] = Y[1]; vgl5[6] = 0.0; vgl5[7] = 2.0; vgl5[8] = 0.0;
vgl1[3] = Y[2]; vgl5[9] = 2.0;
vgl2[1] = 1.0;
vgl3[2] = 1.0;
vgl4[3] = 1.0;
m=4;
double da = 2.0;
for (int32_t a=2 ; a>=0 ; --a) {
double db = 2.0-da;
for (int32_t b=2-a ; b>=0 ; --b) {
const int32_t c = 2 - a - b;
const double dc = 2.0 - da - db;
double xy = pows[0][a+2] * pows[1][b+2];
double yz = pows[1][b+2] * pows[2][c+2];
double xz = pows[0][a+2] * pows[2][c+2];
l[m][0] = a;
l[m][1] = b;
l[m][2] = c;
vgl1[m] = xy * pows[2][c+2];
xy *= dc;
xz *= db;
yz *= da;
vgl2[m] = pows[0][a+1] * yz;
vgl3[m] = pows[1][b+1] * xz;
vgl4[m] = pows[2][c+1] * xy;
vgl5[m] = (da-1.) * pows[0][a] * yz +
(db-1.) * pows[1][b] * xz +
(dc-1.) * pows[2][c] * xy;
db -= 1.0;
++m;
}
da -= 1.0;
} }
m=10;
break; break;
} }
/* /*
case 3: case 3:
{ {
if (ldv < 20) return QMCKL_INVALID_ARG_9; if (ldv < 20) return QMCKL_INVALID_ARG_9;
} }
,*/ ,*/
default: default:
{ {
const int32_t size_max = (lmax+1)*(lmax+2)*(lmax+3)/6; const int32_t size_max = (lmax+1)*(lmax+2)*(lmax+3)/6;
if (ldv < size_max) return QMCKL_INVALID_ARG_9; if (ldv < size_max) return QMCKL_INVALID_ARG_9;
double* const vgl1 = VGL;
double* const vgl2 = VGL + ldv;
double* const vgl3 = VGL + (ldv << 1);
double* const vgl4 = VGL + 3*ldv;
double* const vgl5 = VGL + (ldv << 2);
const double Y[3] = { X[0]-R[0], const double Y[3] = { X[0]-R[0],
X[1]-R[1], X[1]-R[1],
X[2]-R[2] }; X[2]-R[2] };
double pows[3][size_max]; double pows[3][size_max];
for (int32_t i=0 ; i<=2 ; ++i) { for (int32_t i=0 ; i<=2 ; ++i) {
@ -4949,36 +4972,38 @@ assert(0 == test_qmckl_ao_polynomial_vgl(context));
double X[3] = { 1.1, 2.2, 3.3 }; double X[3] = { 1.1, 2.2, 3.3 };
double R[3] = { 0.2, 1.1, 3.0 }; double R[3] = { 0.2, 1.1, 3.0 };
int64_t n; int32_t ldv[4] = {1, 4, 10, 20};
int64_t ldl = 4; for (int32_t ldl=3 ; ldl<5 ; ++ldl) {
int64_t ldv = 24; int64_t n;
int32_t L0[24][ldl]; int32_t L0[24][ldl];
int32_t L1[24][ldl]; int32_t L1[24][ldl];
double VGL0[5][ldv]; printf("ldl=%d\n", ldl);
double VGL1[5][ldv]; for (int32_t lmax=0 ; lmax<=3 ; lmax++) {
memset(&L0[0][0], 0, sizeof(L0)); double VGL0[5][ldv[lmax]];
memset(&L1[0][0], 0, sizeof(L1)); double VGL1[5][ldv[lmax]];
memset(&VGL0[0][0], 0, sizeof(VGL0)); memset(&L0[0][0], 0, sizeof(L0));
memset(&VGL1[0][0], 0, sizeof(VGL1)); memset(&L1[0][0], 0, sizeof(L1));
for (int32_t lmax=0 ; lmax<=3 ; lmax++) { memset(&VGL0[0][0], 0, sizeof(VGL0));
rc = qmckl_ao_polynomial_transp_vgl_doc (context, X, R, lmax, &n, &(L0[0][0]), ldl, &(VGL0[0][0]), ldv); memset(&VGL1[0][0], 0, sizeof(VGL1));
assert (rc == QMCKL_SUCCESS); rc = qmckl_ao_polynomial_transp_vgl_doc (context, X, R, lmax, &n, &(L0[0][0]), ldl, &(VGL0[0][0]), ldv[lmax]);
rc = qmckl_ao_polynomial_transp_vgl_hpc (context, X, R, lmax, &n, &(L1[0][0]), ldl, &(VGL1[0][0]), ldv); assert (rc == QMCKL_SUCCESS);
assert (rc == QMCKL_SUCCESS); rc = qmckl_ao_polynomial_transp_vgl_hpc (context, X, R, lmax, &n, &(L1[0][0]), ldl, &(VGL1[0][0]), ldv[lmax]);
printf("lmax=%d\n", lmax); assert (rc == QMCKL_SUCCESS);
for (int32_t l=0 ; l<n ; ++l) { printf("lmax=%d\n", lmax);
for (int32_t k=0 ; k<3 ; ++k) { for (int32_t l=0 ; l<n ; ++l) {
printf("L[%d][%d] = %d %d\n", l, k, L0[l][k], L1[l][k]); for (int32_t k=0 ; k<3 ; ++k) {
assert( L0[l][k] == L1[l][k] ); printf("L[%d][%d] = %d %d\n", l, k, L0[l][k], L1[l][k]);
} assert( L0[l][k] == L1[l][k] );
} }
}
for (int32_t k=0 ; k<5 ; ++k) { for (int32_t k=0 ; k<5 ; ++k) {
for (int32_t l=0 ; l<n ; ++l) { for (int32_t l=0 ; l<n ; ++l) {
printf("VGL[%d][%d] = %e %e\n", k, l, VGL0[k][l], VGL1[k][l]); printf("VGL[%d][%d] = %e %e\n", k, l, VGL0[k][l], VGL1[k][l]);
assert( VGL0[k][l] == VGL1[k][l] ); assert( VGL0[k][l] == VGL1[k][l] );
} }
} }
}
} }
#+end_src #+end_src
@ -5251,7 +5276,7 @@ qmckl_compute_ao_vgl_hpc (
int64_t n_poly; int64_t n_poly;
const qmckl_exit_code rc = const qmckl_exit_code rc =
qmckl_ao_polynomial_transp_vgl(context, e_coord, n_coord, qmckl_ao_polynomial_transp_vgl_hpc(context, e_coord, n_coord,
nucleus_max_ang_mom[inucl], &n_poly, powers, (int64_t) 3, nucleus_max_ang_mom[inucl], &n_poly, powers, (int64_t) 3,
&(poly_vgl[0][0]), size_max); &(poly_vgl[0][0]), size_max);