160 lines
4.9 KiB
C++
160 lines
4.9 KiB
C++
/********************************************************
|
|
* ██████╗ ██████╗████████╗██╗
|
|
* ██╔════╝ ██╔════╝╚══██╔══╝██║
|
|
* ██║ ███╗██║ ██║ ██║
|
|
* ██║ ██║██║ ██║ ██║
|
|
* ╚██████╔╝╚██████╗ ██║ ███████╗
|
|
* ╚═════╝ ╚═════╝ ╚═╝ ╚══════╝
|
|
* Geophysical Computational Tools & Library (GCTL)
|
|
*
|
|
* Copyright (c) 2022 Yi Zhang (yizhang-geo@zju.edu.cn)
|
|
*
|
|
* GCTL is distributed under a dual licensing scheme. You can redistribute
|
|
* it and/or modify it under the terms of the GNU Lesser General Public
|
|
* License as published by the Free Software Foundation, either version 2
|
|
* of the License, or (at your option) any later version. You should have
|
|
* received a copy of the GNU Lesser General Public License along with this
|
|
* program. If not, see <http://www.gnu.org/licenses/>.
|
|
*
|
|
* If the terms and conditions of the LGPL v.2. would prevent you from using
|
|
* the GCTL, please consider the option to obtain a commercial license for a
|
|
* fee. These licenses are offered by the GCTL's original author. As a rule,
|
|
* licenses are provided "as-is", unlimited in time for a one time fee. Please
|
|
* send corresponding requests to: yizhang-geo@zju.edu.cn. Please do not forget
|
|
* to include some description of your company and the realm of its activities.
|
|
* Also add information on how to contact you by electronic and paper mail.
|
|
******************************************************/
|
|
|
|
#include "gctl/core.h"
|
|
#include "gctl/algorithm.h"
|
|
#include "lcg/lcg.h"
|
|
#include "ctime"
|
|
|
|
#define M 100
|
|
#define N 80
|
|
|
|
// 普通二维数组做核矩阵
|
|
gctl::matrix<double> kernel(M, N, 0.0);
|
|
// 稀疏矩阵为核矩阵
|
|
gctl::spmat<double> sp_kernel(M, N, 0.0);
|
|
// 中间结果数组
|
|
gctl::array<double> tmp_arr(M, 0.0);
|
|
|
|
// 计算核矩阵乘向量的乘积
|
|
void CalAx(void* instance, const lcg_float* x, lcg_float* prod_Ax, const int n_s)
|
|
{
|
|
for (int i = 0; i < M; i++)
|
|
{
|
|
tmp_arr[i] = 0.0;
|
|
for (int j = 0; j < n_s; j++)
|
|
{
|
|
tmp_arr[i] += kernel[i][j] * x[j];
|
|
}
|
|
}
|
|
|
|
for (int j = 0; j < n_s; j++)
|
|
{
|
|
prod_Ax[j] = 0.0;
|
|
for (int i = 0; i < M; i++)
|
|
{
|
|
prod_Ax[j] += kernel[i][j] * tmp_arr[i];
|
|
}
|
|
}
|
|
return;
|
|
}
|
|
|
|
// 计算核矩阵乘向量的乘积
|
|
void CalAx_Spmat(void* instance, const lcg_float* x, lcg_float* prod_Ax, const int n_s)
|
|
{
|
|
// 直接调用稀疏矩阵与向量的乘法
|
|
// 注意第二次为向量乘矩阵 相当于矩阵的转置与向量相乘
|
|
sp_kernel.multiply_vector(x, n_s, tmp_arr.get(), M);
|
|
sp_kernel.multiply_vector(tmp_arr.get(), M, prod_Ax, n_s, gctl::Trans);
|
|
return;
|
|
}
|
|
|
|
//定义共轭梯度监控函数
|
|
int Prog(void* instance, const lcg_float* m, const lcg_float converge, const lcg_para* param, const int n_s, const int k)
|
|
{
|
|
std::clog << "Iteration-times: " << k << "\tconvergence: " << converge << std::endl;
|
|
if (converge > param->epsilon) std::clog << "\033[1A\033[K";
|
|
return 0;
|
|
}
|
|
|
|
int main(int argc, char const *argv[])
|
|
{
|
|
srand(time(0));
|
|
// 添加一些大数
|
|
int tmp_id, tmp_size;
|
|
double tmp_val;
|
|
for (int i = 0; i < M; i++)
|
|
{
|
|
tmp_size = gctl::random(25, 35);
|
|
for (int j = 0; j < tmp_size; j++)
|
|
{
|
|
tmp_id = gctl::random(0, N);
|
|
tmp_val = gctl::random(-10.0, 10.0);
|
|
|
|
kernel[i][tmp_id] = tmp_val;
|
|
sp_kernel.insert(i, tmp_id, tmp_val);
|
|
}
|
|
}
|
|
|
|
// 生成一组正演解
|
|
gctl::array<double> fm(N);
|
|
for (int i = 0; i < N; i++)
|
|
{
|
|
fm[i] = gctl::random(1.0, 2.0);
|
|
}
|
|
|
|
// 计算共轭梯度B项
|
|
gctl::array<double> B(N);
|
|
sp_kernel.multiply_vector(fm.get(), N, tmp_arr.get(), M);
|
|
sp_kernel.multiply_vector(tmp_arr.get(), M, B.get(), N, gctl::Trans);
|
|
/*
|
|
for (int i = 0; i < M; i++)
|
|
{
|
|
tmp_arr[i] = 0.0;
|
|
for (int j = 0; j < N; j++)
|
|
{
|
|
tmp_arr[i] += kernel[i][j]*fm[j];
|
|
}
|
|
}
|
|
|
|
for (int j = 0; j < N; j++)
|
|
{
|
|
B[j] = 0.0;
|
|
for (int i = 0; i < M; i++)
|
|
{
|
|
B[j] += kernel[i][j]*tmp_arr[i];
|
|
}
|
|
}
|
|
*/
|
|
|
|
/********************准备工作完成************************/
|
|
lcg_para self_para = lcg_default_parameters();
|
|
self_para.max_iterations = 1000;
|
|
self_para.epsilon = 1e-10;
|
|
|
|
// 声明两组解
|
|
gctl::array<double> m(N, 0.0);
|
|
gctl::array<double> m_sp(N, 0.0);
|
|
|
|
clock_t start = clock();
|
|
int ret = lcg_solver(CalAx, Prog, m.get(), B.get(), N, &self_para, NULL, LCG_CG);
|
|
clock_t end = clock();
|
|
if (ret < 0) lcg_error_str(ret);
|
|
std::cout << "array2d's time: " << 1000.0*(end - start)/(double)CLOCKS_PER_SEC << " ms" << std::endl;
|
|
|
|
start = clock();
|
|
ret = lcg_solver(CalAx_Spmat, Prog, m_sp.get(), B.get(), N, &self_para, NULL, LCG_CG);
|
|
if (ret < 0) lcg_error_str(ret);
|
|
end = clock();
|
|
std::cout << "spmat's time: " << 1000.0*(end - start)/(double)CLOCKS_PER_SEC << " ms" << std::endl;
|
|
|
|
for (int i = 0; i < N; i++)
|
|
{
|
|
std::cout << fm[i] << " " << m[i] << " " << m_sp[i] << std::endl;
|
|
}
|
|
return 0;
|
|
} |