Computation of Lyapunov exponents

This commit is contained in:
Ian Jauslin 2024-02-19 19:01:31 -05:00
parent 89601d19d1
commit b31fab8eae
2 changed files with 215 additions and 3 deletions

View File

@ -16,10 +16,134 @@ limitations under the License.
#include "constants.cpp" #include "constants.cpp"
#include "lyapunov.h" #include "lyapunov.h"
#include <cblas.h> #include <openblas64/cblas.h>
#include <openblas64/lapacke.h>
#include <math.h>
#include <stdlib.h>
#define MATSIZE (K1*(2*(K2+1)+K2)) #define MATSIZE (K1*(2*(K2+1)+K2))
// compute Lyapunov exponents
int lyapunov(
int K1,
int K2,
int N1,
int N2,
double final_time,
double lyapunov_reset,
double nu,
double D_epsilon,
double delta,
double L,
double adaptive_tolerance,
double adaptive_factor,
double max_delta,
unsigned int adaptive_norm,
_Complex double* u0,
_Complex double* g,
bool irreversible,
bool keep_en_cst,
double target_en,
unsigned int algorithm,
double starting_time,
unsigned int nthreads
){
double* lyapunov;
double* lyapunov_avg;
_Complex double* Du_prod;
_Complex double* Du;
_Complex double* u;
_Complex double* prevu;
_Complex double* tmp1;
_Complex double* tmp2;
_Complex double* tmp3;
_Complex double* tmp4;
_Complex double* tmp5;
_Complex double* tmp6;
_Complex double* tmp7;
_Complex double* tmp8;
_Complex double* tmp9;
_Complex double* tmp10;
double time;
fft_vect fft1;
fft_vect fft2;
fft_vect ifft;
// period
// add 0.1 to ensure proper rounding
uint64_t n=(uint64_t)((starting_time-fmod(starting_time, lyapunov_reset))/lyapunov_reset+0.1);
lyapunov_init_tmps(&lyapunov, &lyapunov_avg, &Du_prod, &Du, &u, &prevu, &tmp1, &tmp2, &tmp3, &tmp4, &tmp5, &tmp6, &tmp7, &tmp8, &tmp9, &tmp10, &fft1, &fft2, &ifft, K1, K2, N1, N2, nthreads, algorithm);
// copy initial condition
copy_u(u, u0, K1, K2);
// store step (useful for adaptive step methods
double prev_step=delta;
double step=delta;
double next_step=step;
const _Complex double one=1;
const _Complex double zero=0;
int i;
// init average
for (i=0; i<MATSIZE; i++){
lyapunov_avg=0;
}
// save times at which Lyapunov exponents are computed
double prevtime=starting_time;
// iterate
time=starting_time;
while(final_time<0 || time<final_time){
// copy before step
copy_u(prevu, u, K1, K2);
prev_step=step;
ns_step(algorithm, u, K1, K2, N1, N2, nu, &step, &next_step, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, starting_time, fft1, fft2, ifft, &tmp1, &tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, irreversible, keep_en_cst, target_en);
// compute Jacobian
// do not touch tmp1, it might be used in the next step
ns_D_step(Du, prevu, u, K1, K2, N1, N2, nu, D_epsilon, prev_step, algorithm, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, fft1, fft2, ifft, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, irreversible, keep_en_cst, target_en);
// multiply Jacobian
// switch to column major order, so transpose Du
cblas_zgemm(CblasColMajor, CblasNoTrans, CblasTrans, MATSIZE, MATSIZE, MATSIZE, &one, Du_prod, MATSIZE, Du, MATSIZE, &zero, tmp10, MATSIZE);
// copy back to Du_prod
_Complex double* move=tmp10;
tmp10=Du_prod;
Du_prod=move;
// increment time
time+=step;
// reset Jacobian
if(time>(n+1)*lyapunov_reset){
n++;
// QR decomposition
// do not touch tmp1, it might be used in the next step
LAPACKE_zgerqf(LAPACK_COL_MAJOR, MATSIZE, MATSIZE, Du_prod, MATSIZE, tmp2);
// extract eigenvalues (diagonal elements of Du_prod)
for(i=0; i<MATSIZE; i++){
lyapunov[i]=log(cabs(Du_prod[i*MATSIZE+i]))/(time-prevtime);
// add to average
lyapunov_avg[i]=lyapunov_avg[i]*prevtime/time+lyapunov[i]*(time-prevtime)/time;
}
// set Du_prod to Q:
LAPACKE_zungrq(LAPACK_COL_MAJOR, MATSIZE, MATSIZE, MATSIZE, Du_prod, MATSIZE, tmp2);
// reset prevtime
prevtime=time;
}
}
lyapunov_free_tmps(lyapunov, lyapunov_avg, Du_prod, Du, u, prevu, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, tmp10, fft1, fft2, ifft, algorithm);
return(0);
}
// Jacobian of u_n -> u_{n+1} // Jacobian of u_n -> u_{n+1}
int ns_D_step( int ns_D_step(
_Complex double* out, _Complex double* out,
@ -77,7 +201,8 @@ int ns_D_step(
ns_step(algorithm, tmp1, K1, K2, N1, N2, nu, &step, &next_step, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, time, fft1, fft2, ifft, &tmp2, &tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, irreversible, keep_en_cst, target_en); ns_step(algorithm, tmp1, K1, K2, N1, N2, nu, &step, &next_step, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, time, fft1, fft2, ifft, &tmp2, &tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, irreversible, keep_en_cst, target_en);
// compute derivative // compute derivative
for (i=0;i<MATSIZE;i++){ for (i=0;i<MATSIZE;i++){
out[i]=(tmp1[i]-un_next[i])/epsilon; // use row major order
out[klookup_sym(lx,ly,K2)*MATSIZE+i]=(tmp1[i]-un_next[i])/epsilon;
} }
// derivative in the imaginary direction // derivative in the imaginary direction
@ -96,7 +221,8 @@ int ns_D_step(
ns_step(algorithm, tmp1, K1, K2, N1, N2, nu, &step, &next_step, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, time, fft1, fft2, ifft, &tmp2, &tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, irreversible, keep_en_cst, target_en); ns_step(algorithm, tmp1, K1, K2, N1, N2, nu, &step, &next_step, adaptive_tolerance, adaptive_factor, max_delta, adaptive_norm, L, g, time, time, fft1, fft2, ifft, &tmp2, &tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, irreversible, keep_en_cst, target_en);
// compute derivative // compute derivative
for (i=0;i<MATSIZE;i++){ for (i=0;i<MATSIZE;i++){
out[i]+=-(tmp1[i]-un_next[i])/epsilon*I; // use row major order
out[klookup_sym(lx,ly,K2)*MATSIZE+i]=-(tmp1[i]-un_next[i])/epsilon*I;
} }
} }
} }
@ -105,3 +231,79 @@ int ns_D_step(
} }
int lyapunov_init_tmps(
double ** lyapunov,
double ** lyapunov_avg,
_Complex double ** Du_prod,
_Complex double ** Du,
_Complex double ** u,
_Complex double ** prevu,
_Complex double ** tmp1,
_Complex double ** tmp2,
_Complex double ** tmp3,
_Complex double ** tmp4,
_Complex double ** tmp5,
_Complex double ** tmp6,
_Complex double ** tmp7,
_Complex double ** tmp8,
_Complex double ** tmp9,
_Complex double ** tmp10,
fft_vect* fft1,
fft_vect* fft2,
fft_vect* ifft,
int K1,
int K2,
int N1,
int N2,
unsigned int nthreads,
unsigned int algorithm
){
ns_init_tmps(u, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, fft1, fft2, ifft, K1, K2, N1, N2, nthreads, algorithm);
*lyapunov=calloc(sizeof(double),MATSIZE);
*lyapunov_avg=calloc(sizeof(double),MATSIZE);
*Du_prod=calloc(sizeof(_Complex double),MATSIZE*MATSIZE);
*Du=calloc(sizeof(_Complex double),MATSIZE*MATSIZE);
*prevu=calloc(sizeof(_Complex double),MATSIZE);
*tmp1=calloc(sizeof(_Complex double),MATSIZE);
*tmp2=calloc(sizeof(_Complex double),MATSIZE);
*tmp10=calloc(sizeof(_Complex double),MATSIZE*MATSIZE);
return(0);
}
// release vectors
int lyapunov_free_tmps(
double* lyapunov,
double* lyapunov_avg,
_Complex double* Du_prod,
_Complex double* Du,
_Complex double* u,
_Complex double* prevu,
_Complex double* tmp1,
_Complex double* tmp2,
_Complex double* tmp3,
_Complex double* tmp4,
_Complex double* tmp5,
_Complex double* tmp6,
_Complex double* tmp7,
_Complex double* tmp8,
_Complex double* tmp9,
_Complex double* tmp10,
fft_vect fft1,
fft_vect fft2,
fft_vect ifft,
unsigned int algorithm
){
free(tmp10);
free(tmp2);
free(tmp1);
free(prevu);
free(Du);
free(Du_prod);
free(lyapunov_avg);
free(lyapunov);
ns_free_tmps(u, tmp3, tmp4, tmp5, tmp6, tmp7, tmp8, tmp9, fft1, fft2, ifft, algorithm);
return(0);
}

View File

@ -19,7 +19,17 @@ limitations under the License.
#include "navier-stokes.h" #include "navier-stokes.h"
// compute Lyapunov exponents
int lyapunov( int K1, int K2, int N1, int N2, double final_time, double lyapunov_reset, double nu, double D_epsilon, double delta, double L, double adaptive_tolerance, double adaptive_factor, double max_delta, unsigned int adaptive_norm, _Complex double* u0, _Complex double* g, bool irreversible, bool keep_en_cst, double target_en, unsigned int algorithm, double starting_time, unsigned int nthreads);
// Jacobian of step
int ns_D_step( _Complex double* out, _Complex double* un, _Complex double* un_next, int K1, int K2, int N1, int N2, double nu, double epsilon, double delta, unsigned int algorithm, double adaptive_tolerance, double adaptive_factor, double max_delta, unsigned int adaptive_norm, double L, _Complex double* g, double time, fft_vect fft1, fft_vect fft2, fft_vect ifft, _Complex double* tmp1, _Complex double* tmp2, _Complex double* tmp3, _Complex double* tmp4, _Complex double* tmp5, _Complex double* tmp6, _Complex double* tmp7, _Complex double* tmp8, bool irreversible, bool keep_en_cst, double target_en); int ns_D_step( _Complex double* out, _Complex double* un, _Complex double* un_next, int K1, int K2, int N1, int N2, double nu, double epsilon, double delta, unsigned int algorithm, double adaptive_tolerance, double adaptive_factor, double max_delta, unsigned int adaptive_norm, double L, _Complex double* g, double time, fft_vect fft1, fft_vect fft2, fft_vect ifft, _Complex double* tmp1, _Complex double* tmp2, _Complex double* tmp3, _Complex double* tmp4, _Complex double* tmp5, _Complex double* tmp6, _Complex double* tmp7, _Complex double* tmp8, bool irreversible, bool keep_en_cst, double target_en);
// init vectors
int lyapunov_init_tmps(double** lyapunov, double** lyapunov_avg, _Complex double ** Du_prod, _Complex double ** Du, _Complex double ** u, _Complex double ** prevu, _Complex double ** tmp1, _Complex double ** tmp2, _Complex double ** tmp3, _Complex double ** tmp4, _Complex double ** tmp5, _Complex double ** tmp6, _Complex double ** tmp7, _Complex double ** tmp8, _Complex double ** tmp9, _Complex double** tmp10, fft_vect* fft1, fft_vect* fft2, fft_vect* ifft, int K1, int K2, int N1, int N2, unsigned int nthreads, unsigned int algorithm);
// release vectors
int lyapunov_free_tmps(double* lyapunov, double* lyapunov_avg, _Complex double* Du_prod, _Complex double* Du, _Complex double* u, _Complex double* prevu, _Complex double* tmp1, _Complex double* tmp2, _Complex double* tmp3, _Complex double* tmp4, _Complex double* tmp5, _Complex double* tmp6, _Complex double* tmp7, _Complex double* tmp8, _Complex double* tmp9, _Complex double* tmp10, fft_vect fft1, fft_vect fft2, fft_vect ifft, unsigned int algorithm);
#endif #endif