-
Notifications
You must be signed in to change notification settings - Fork 106
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Boundserror with multiple right hand sides #318
Comments
I talked a bit with @fredrikekre . Iterative solvers might not be suited for multiple RHS since it is unclear which residual to check? |
A trivial version would be to just run the exact equivalent of [cg(A, b[:, 1]), cg(A, b[:, 2]), ..., cg(A, b[:, n])] with an optimization where matrix multiplications act on a block vector, so that you have a higher flop/memop ratio. But you need to keep in mind vectors converge at different speeds, so the matrix multiplication should only happen on the block that's not converged (you can split it like |
Do you think there would be an advantage of doing
where over just using one rhs at the time:
Lets assume the vectors will converge with roughly the same speed. |
I get an error when not allocating
x
when I have multiple RHS.I guess cg(A,b) should also work?
The text was updated successfully, but these errors were encountered: