columbia-computational-x-lab / diffsmat Goto Github PK

Thank you for making the source code for your paper available. I was wondering if you could provide more examples demonstrating how different geometries can be used during an optimization run. For instance, is it possible for you to additionally share the code that generates the results in Fig 7 of the paper?

When I tried to compile DiffSMat using Eigen 3.3.9, I received compilation errors regarding general_matrix_matrix_product calls in DiffSMat/src/dcomplex/DComplexMatrixFunc.hpp.

It seems like the following issue in Eigen lead to changes in how general_matrix_matrix_product is called. There are now 13 parameters in struct definition, and run function has an extra parameter Index resIncr.
https://gitlab.com/libeigen/eigen/-/issues/1741

The following lines therefore need to be modified.
https://github.com/Columbia-Computational-X-Lab/DiffSMat/blob/main/src/dcomplex/DComplexMatrixFunc.hpp#L218
https://github.com/Columbia-Computational-X-Lab/DiffSMat/blob/main/src/dcomplex/DComplexMatrixFunc.hpp#L225
https://github.com/Columbia-Computational-X-Lab/DiffSMat/blob/main/src/dcomplex/DComplexMatrixFunc.hpp#L235
https://github.com/Columbia-Computational-X-Lab/DiffSMat/blob/main/src/dcomplex/DComplexMatrixFunc.hpp#L242