You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I use pylops for solving large linear systems, and to test preconditioners.
Since pylops.optimization.solver.cg is very close to scipy.sparse.linalg.cg, I wondered if one could implement this feature, where the preconditioner is given as a linear operator which approximates the inverse of the original matrix (scipy's cg)
I have already implemented it as an new solver, if you are interested, I can prepare a PR.
Thank you
The text was updated successfully, but these errors were encountered:
Hello,
Indeed the two solvers should be identical. The only reason we implemented cg was to have a solver compatible with cupy arrays (scipy one is not). If you have this feature and can add it to our cg in a similar fashion to what scipy does, that would be great. Feel free to make a PR at any time :)
Hi,
I use pylops for solving large linear systems, and to test preconditioners.
Since
pylops.optimization.solver.cg
is very close toscipy.sparse.linalg.cg
, I wondered if one could implement this feature, where the preconditioner is given as a linear operator which approximates the inverse of the original matrix (scipy's cg)I have already implemented it as an new solver, if you are interested, I can prepare a PR.
Thank you
The text was updated successfully, but these errors were encountered: