Add support for PETSc solvers #659
Closed
Conversation
This file contains 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
Serial only. Must stipulate bandwidth in all uses. Addresses usnistgov#644
There was a lot of redundant code between the specified solver cases and the guessing solver cases. These treatements have been consolidated and simplified. Easier to rearrange order of solvers to try and to add in new solvers. Addresses usnistgov#644
Old `parallelComm` made heavy assumptions about availability of Trilinos. Comms (particularly the parallel ones) are intimately tied to the solver in use, so they should be instantiated with the solver. No matter what, we need mpi4py, so use it (and only it) to determine number of processes in order to pick a solver. Addresses usnistgov#644
Based on trilinos' comms Addresses usnistgov#644
PETSc seems pickier about matrices with nothing on the diagonal than the other solvers. This `DummySolver` is an attempt to work around that, although I'm not clear what the `DummySolver` is attempting to resolve in the first place. Addresses usnistgov#644
Not used and no longer importable at this stage. Addresses usnistgov#644
Matrices and vectors have already been converted by the time we get to `_solve_()`. Addresses usnistgov#644
There is no point in passing the iterations and tolerance to `PETSc.KSP()`. It is ony used to precondition. We need to handle iterations and tolerance ourselves, just like for the other LU solvers. Addresses usnistgov#644
Incomplete transition to parallel. Works for some (`examples/diffusion/mesh1D.py`, `examples/diffusion/mesh20x20.py`), but not others (`examples/diffusion/circle.py`). Addresses usnistgov#644
PETSc's LinearLUSolver presently doesn't converge, causing this example to run forever. Addresses usnistgov#644
Matrix multiplication did not properly deal with parallel PETSc vectors or with 32 bit indices
All (expected) tests pass
numerix routines are no longer imported globally. Horrible construction of chemotaxis tests meant we didn't catch these two.
More instructive display of parallel conditions, including errors
The need for a different `__mul__` for `_PETScMatrix` and `_PETScMeshMatrix` got lost in the mix. This is an attempt to get the right differentiation, based loosely on TrilinosMatrix.
Note: PETSc Matrix market export is broken. Patch from https://bitbucket.org/jeguyer/petsc/branch/fix-matrix-market-export is required.
PETSc vectors don't support addition with NumPy arrays. Epetra vectors do support it, but things also seem to work if everything is cast to NumPy arrays.
Ensure that result of a matrix * matrix multiplication doesn't downcast
|
I couldn't find any typos in the text you've written. All the links seemed to work as well. The only thing is the graph. I would make clear what the numbers mean in the legend. In the caption write something like "Scaling behaviour of different solver packages with both varying tasks and varying threads. The number of tasks (processes) varies along the x-axis and the number of threads varies via the line style (see legend)." The bullet points are an explanation for the behavior in the graph, but you don't describe the behavior. For example, I think you should say "Trilinos with two threads is faster than with one thread until the number of tasks reaches 10. We don't fully understand the reason for this" etc. I find it hard to interpret graphs like that without a narrative. |
tkphd
approved these changes
Jan 22, 2020
Respons to [feedback from @wd15](usnistgov#659 (comment))
Response to [@tkphd comments](usnistgov#659 (review))
|
Looks good now. |
wd15
approved these changes
Jan 22, 2020
This was referenced Jan 23, 2020
Closed
Gmsh 4.0 doesn't partition correctly
[Ghosts are broken](https://gitlab.onelab.info/gmsh/gmsh/issues/733), but export them anyway.
Inexplicable error:
convert: improper image header `/root/project/documentation/_build/latex/.doctrees/images/dae02450438e07c40e1e06ac9a26251fb6896047/project_thin_badge.gif' @ error/gif.c/ReadGIFImage/1028.
convert: no images defined `/root/project/documentation/_build/latex/.doctrees/images/project_thin_badge.png' @ error/convert.c/ConvertImageCommand/3273.
Gmsh refinement seems to stall out
|
Superceded by #701 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Addresses #387
TODO:
This situation is unclear and needs more analysis.
This situation is unclear and needs more analysis.
nproc> 2Failures in
Doctest: fipy.meshes.gmshMesh.Gmsh2DDoctest: examples.diffusion.nthOrder.input4thOrder1DDoctest: fipy.meshes.representations.gridRepresentation._GridRepresentation._testDoctest: examples.levelSet.distanceFunction.square(due to skfmm raises exception on empty mesh #684, which is beyond the scope of this PR)environment.yml/requirements.txt(where are these?!? Ah, inbinder/.)