feat: [linalg] add iterative solvers #994
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…r workspace sizes
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
Contributor
Author
|
@jvdp1 thanks for your reviews. I think I managed to address all of them. Let me know your thoughts. |
jvdp1
approved these changes
Sep 19, 2025
Co-authored-by: Jeremie Vandenplas <[email protected]>
Co-authored-by: Jeremie Vandenplas <[email protected]>
jvdp1
reviewed
Sep 23, 2025
Contributor
Author
|
With no more comments I'll go ahead and merge this. |
4 tasks
Mahmood-Sinan
pushed a commit
to Mahmood-Sinan/stdlib
that referenced
this pull request
Dec 17, 2025
* start iterative solvers * simplify workspace * add pccg solver and example * complete cg with dirichlet flag, add example, fix di filter * add other sparse matrices * add example for custom solver extending the generic interface * small simplifications for working data * make default inner_product point to a default dot_product add enum for workspace sizes * make preconditionner a linop * use facility size * Add a customizable logger facility, change linop matvec to apply * change internal procedure names for custom example * refactor to remove hard dependency on dirichlet BCs * add forward/backward solvers for preconditionning * fix solve forward/backward * fix colum index * add preconditionners * fix cmake * add factorizations * backward solve to use Lt marix * start unit testing * review csr factorization * change name generic for kernel * shorten factorization procedures names * change precond ldl name * rename to pcg * start working on the documentation * clean docstrings * add cg and pcg tests * Update example/linalg/example_solve_cg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * add _type suffix * reduce PR scope * rename wksp size constants * rename constant * change signature of linop apply to matvec and add operator input argument * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update test/linalg/test_linalg_solve_iterative.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * Update test/linalg/test_linalg_solve_iterative.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * Update src/stdlib_linalg_iterative_solvers_cg.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * Update src/stdlib_linalg_iterative_solvers_cg.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_cg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_cg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_cg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update example/linalg/example_solve_pcg.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update src/stdlib_linalg_iterative_solvers_cg.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * small fixes * add documentation information * forgotten attribute in doc * use optval * explicit use of private/public * Update doc/specs/stdlib_linalg_iterative_solvers.md Co-authored-by: Jeremie Vandenplas <[email protected]> * rename with stdlib_ prefix * Update example/linalg/example_solve_custom.f90 Co-authored-by: Jeremie Vandenplas <[email protected]> * Update src/stdlib_linalg_iterative_solvers_pcg.fypp Co-authored-by: Jeremie Vandenplas <[email protected]> * replace tol by rtol and atol for convergence test * change rtol and atol defaults --------- Co-authored-by: Jeremie Vandenplas <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR introduces the base for iterative solvers.
Two methods are proposed:
Each method is made public with two public interface flavors:
stdlib_solve_<method>_kernel: All arguments are mandatory (no optionals/no internal allocations), it contains the methods steps. The linear system (and preconditioner) is defined through a public DTstdlib_linopwhich enables to extend two key procedures:matvecequivalent to a matrix-vector product andinner_productequivalent to the dot_product. This is the interface recommended to extend the method when dealing with a matrix type not available instdlibor when working in distributed-memory frameworks for which the matvec, dot_product and factorization steps need to be adapted to account for parallel synchronization.stdlib_solve_<method>: API with optional arguments, the linear system can be defined with dense orCSR_<>_typematrices. For the PCG, the following preconditioners are available:none(identity),jacobi(1/diagonal). It internally usesstdlib_solve_<method>_kernel.