search

GridOPTICS / GridPACK

https://www.gridpack.org/
44 stars 22 forks source link

Add support for real matrices to generalized matvec interface #166

Open bjpalmer opened 1 year ago

bjpalmer commented 1 year ago

Current implementation of generalized matrix-vector interface only supports complex matrices. We need to add support for real matrices as well.

abhyshr commented 1 year ago

This one also is for the EMT application so we can leave it out for the release.

bjpalmer commented 1 year ago

This one may boil down to mostly copy/paste. Could be pretty quick.

abhyshr commented 1 year ago

I have this working in my EMT branch. This was not an easy copy/paste fix since changes made to the mapper also need to set to the math solvers - some of which do not interfaces with real matrices (DAE solver for instance was one math solver that I had to update).

wperkins commented 1 year ago

(DAE solver for instance was one math solver that I had to update).

Are these code changes checked in? I would like to see why this was necessary.

abhyshr commented 1 year ago

Yes, to the fix/testing branch.

abhyshr commented 1 year ago

https://github.com/GridOPTICS/GridPACK/commit/1e6351540daf49ef324bc067d49e14a2f24a935f

abhyshr commented 1 year ago

@wperkins : I did not read your question correctly. The fixes for the memory crash are in fix/testing branch. The changes to the DAE solver are in feature/emt branch.

wperkins commented 1 year ago

@wperkins : I did not read your question correctly. The fixes for the memory crash are in fix/testing branch. The changes to the DAE solver are in feature/emt branch.

My unspecific question was about the DAESolver changes in feature/emt. I also caught those in fix/testing, but we handled them a little differently. Please see my comments there.

I'll try out the dynamic simulation fix. Thanks.