Hi Ralf,

Thanks for your email and happy to assist in whatever way I can. I've reached out to Dr Xiaoyei Li of Lawrence Berkley who appears to be the principal maintainer of the code base with a request to host the _mt code on github. Given the other two versions are already hosted I can't see that being an issue. I've also queried whether developments are done in sync and if so, if there is a pending patching for mt in the similar fashion to sequential.

I will revert with further information as I receive it.

Cheers,
Neil

On Mon, 24 May 2021 at 04:33, Ralf Gommers <ralf.gommers@gmail.com> wrote:


On Fri, May 21, 2021 at 2:11 AM Neil Godber <njgodber@gmail.com> wrote:
Hi dev-list,

Further last, realised that I failed to link to the issue itself despite linking to virtually everything else:


Cheers,
Neil

On Fri, 21 May 2021 at 10:04, Neil Godber <njgodber@gmail.com> wrote:
Hi dev-list,

Thanks for the proposal Neil.


I've logged an enhancement request on github to bring the parallel versions of SuperLU into scipy. As requested I'm circulating the enhancement here to solicit comment/feedback/engagement. 

Content of request is as follows (with light editing and links):

"Is your feature request related to a problem? Please describe.

The direct solution of sparse matrices is a common problem that arises across many domains. Current scipy.sparse.linalg provides spsolve and splu to solve such classes of problems. Neither solution appears to utilise more than one or two threads leaving a lot of the available compute on modern machines unused. Other solutions (MUMPS, Pardiso) exploit multicore processors but are either minimally supported and/or locked behind proprietary libraries.

Describe the solution you'd like
Fortunately SuperLU has two multithreaded implementations superLU_mt and superLU_dt, both actively maintained (SuperLU: Home Page (nersc.gov)). Ideally SciPy would replace the current sequential implementation with one of the parallel implementations to yield large performance improvements.

The relevant one seems to be superLU_mt. The `_dt` flavor is for distributed computing, which is out of scope for SciPy. superLU_mt says it has both pthreads and OpenMP interfaces - we'd prefer pthreads (and actually forbid OpenMP within SciPy).

The last release of superLU_mt is v3.1, from March 2016. Our current copy of SuperLU is 5.2.1, from May 2016 - but then there's a bunch of patches applied to it, and then https://github.com/scipy/scipy/pull/12243 did a sync with SuperLU master in May 2020. So it would be necessary to figure out if we can get superLU_mt from upstream master as well, and that we don't lose bug fixes we applied since 2016. Could you look into that Neil? And do you know if the parallel and sequential versions are developed in sync or not?

Cheers,
Ralf


Describe alternatives you've considered
pyMKL (Pardiso dwfmarchant/pyMKL: Python wrappers to Intel MKL routines (github.com)) - MKL with issues associated with that for non Intel processors, unmaintained relatively to scipy, PyTrilinos (implements all three superLU types but only available on osx and linux PyTrilinos | Trilinos), petsc (petsc4py PETSc / petsc · GitLab) (implements all three superLU types but only available on osx and linux). All are relatively inaccessible compared to ubiquity and accessibility of scipy.

Additional context (e.g. screenshots)
Anecdotally large number of projects use various non scipy sparse solvers to achieve high level of performance. This comes with considerable time and maintenance investment. It would be fantastic for scipy to provide a 'performance competitive' sparse direct solver out of the hood as it does for dense matrices."


Cheers,

Neil

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev@python.org
https://mail.python.org/mailman/listinfo/scipy-dev
_______________________________________________
SciPy-Dev mailing list
SciPy-Dev@python.org
https://mail.python.org/mailman/listinfo/scipy-dev