Test with ESMF 8.1

[huard]

A new release of ESMF is available. We should

• confirm xESMF works, and
• identify new features we want to expose.

https://github.com/esmf-org/esmf/releases/tag/ESMF_8_1_0

[Plantain]

Just from some trivial testing with our own workloads, it seems considerably slower than 8.0.1.

Could we have MOAB exposed through xESMF to try?
The integration of the MOAB mesh library, developed by the U.S. Department of Energy, into ESMF, has reached a milestone that supports application-level testing. All users of ESMF are encouraged to begin testing their applications with the MOAB option turned ON. Preliminary results indicate that MOAB provides improved performance and scaling, and reduced memory footprint. Applications seeking to go to very high resolution grids should benefit from this new capability. The MOAB mesh feature is still OFF by default in this release, but can easily be turned ON from the application level, any time during the run. The available regridding features supported by the MOAB based implementation are clearly listed in the detailed release notes.

[huard]

Is ESMF_MeshSetMOAB() exposed through ESMPy?

My guess is that you'd first need to include it here: https://github.com/esmf-org/esmf/blob/develop/src/addon/ESMPy/src/ESMF/interface/cbindings.py

[rokuingh]

@huard @Plantain we have created a feature request for this under #37.
@Plantain we are optimizing regridding performance in this development cycle. Do you know a specific routine or area of the library that is responsible for the slow down that you have observed between 8.0.1 and 8.1.0? We did not knowingly make any changes that would cause this behaviour, and we would like to identify and fix this performance issue for the patch release 8.1.1 if possible.

[Plantain]

Is that a private repo? I can't see it.

…

On Thu, Apr 22, 2021 at 12:09 PM rokuingh ***@***.***> wrote: @huard <https://github.com/huard> @Plantain <https://github.com/Plantain> we have created a feature request for this under #37 <https://github.com/esmf-org/esmf-support/issues/37>. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#82 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACODF32OZZSJAZ4SPVD4P3TJ6AM5ANCNFSM42B6MWCQ> .

[rokuingh]

Oops, my mistake, it is. We are just moving our issues over to github and still have it set to private. Well, for now just know that we are aware of the issue. If you could let us know what the urgency is on this request it will help us to prioritize it for the upcoming release cycle.

[Plantain]

Well, I don't know about urgent but I think it should be a hard block on making MOAB the default if all the users of ESMF via ESMPy have not been able to test their workflows in advance?

…

On Thu, Apr 22, 2021 at 12:19 PM rokuingh ***@***.***> wrote: Oops, my mistake, it is. We are just moving our issues over to github and still have it set to private. Well, for now just know that we are aware of the issue. If you could let us know what the urgency is on this request it will help us to prioritize it for the upcoming release cycle. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#82 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACODF5CCGF3SIXGY3RB5BDTJ6BU7ANCNFSM42B6MWCQ> .

[rokuingh]

agreed. we don't plan to have MOAB be the default backend to ESMF_Mesh until 8.3.0, two releases from now, about a year out at best guess. This ticket is just to add a flag to allow a user to toggle the MOAB mesh functionality.

[huard]

Thanks @rokuingh for the info and timeline. I was wondering how to submit feature requests to ESMF.

I've opened a ticket to benchmark performance (#91).

[rokuingh]

A feature request can be submitted with an email to esmf_support@ucar.edu.

[huard]

xESMF now tested with a range of ESMF versions, see #230