Hi there, I'm a newbie, could you clarify this:
in .h:

Code: Select all

SparseMatrix<double> K1;
SparseMatrix<double> K2;
SparseMatrix<double> K3;
vector< SparseMatrix<double> > mA //define vector of SparseMatrix objs


in .cpp

Code: Select all

mA.reserve(3);
mA.push_back(K1);
mA.push_back(K2);
mA.push_back(K3);


Since the Sparse Matrices could be huge, is it better to allocate memory and use a vector of pointers to sparse matrix:

Code: Select all

vector< SparseMatrix<double> *> mA //


How do you think? Is the last code useless?

thanks

Hi,

I see two options, 1) you push empty SparseMatrix and directly work with them, 2) you create individual SparseMatrix and then only store a vector of pointers.

ggael wrote:Hi,

I see two options, 1) you push empty SparseMatrix and directly work with them, 2) you create individual SparseMatrix and then only store a vector of pointers.

yes, the code is not complete.
in the last option i would write:

Code: Select all

ma.push_back(new SparseMatrix<double>) //not correct i guess


I don`t understand how Eigen handles the memory.

I meant that if all you want is to create a std::vector of sparse matrices, and not aggregates existing matrices into a std::vector, then use option 1, otherwise use option 2.

It's clear for me.
Thanks