Re: [eigen] Eigen Internals

[Mayank Agarwal <mayank007.win@xxxxxxxxx>]

Hi,

Thanks a lot. I am new to Eigen and have been going through documentation and all.

I want to implement symmetric matrices class in Eigen, for which I guess knowing the internals

and how Eigen works is required. I am not sure which part to learn and what parts are there.

I just want to learn enough so that I know what my approach should be to create such a working

class in Eigen.

Thank you,

Mayank

‌

On Thu, Mar 8, 2018 at 10:14 PM, Gael Guennebaud <gael.guennebaud@xxxxxxxxx> wrote:

Hi,

that's a though question! Which part do you want to learn?

Anyhow, a good start could be to run step by step in a debugger the following code:

#define EIGEN_DONT_VECTORIZE

#include <Eigen/Core>

using namespace Eigen;

int main() {

  int n = 5;

  VectorXd A(n), B(n), C(n);

  A = B+C;

}

This will show you what is returned by A+B, and how this simple _expression_ is evaluated. You can comment #define  EIGEN_DONT_VECTORIZE  to see how explicit vectorization is accomplished or try more fancy expressions.

gael

On Wed, Mar 7, 2018 at 8:07 PM, Mayank Agarwal <mayank007.win@xxxxxxxxx> wrote:

Hi,

I have been going through the core Eigen folder. I am really confused about the starting point of it all as from which header file I should start understanding the Eigen internals and function definitions.

Thank You,

Mayank