Difference between revisions of "Teleconference 2021-03-24"
Line 1: | Line 1: | ||
− | Attendees: | + | Attendees: Ken Moreland (SNL), Nick Davis (SNL), Ollie Lo (LANL), Abhishek Yenpure (UO), Hank Childs (UO), Berk Geveci (Kitware), Sujin Philip (Kitware), Silvio Rizzi (ANL), David Pugmire (ORNL), Matt Larsen (LLNL), Vicente Bolea (Kitware), Terry Turton (LANL) |
== ECP Updates == | == ECP Updates == | ||
Line 8: | Line 8: | ||
** Expected to be available starting around mid May | ** Expected to be available starting around mid May | ||
* New ALCF early access machine: Arcticus | * New ALCF early access machine: Arcticus | ||
− | ** Available now | + | ** Available now |
+ | ** If you have an existing JLSE account, you should now have one on Articus. | ||
* Security changes to ECP confluence | * Security changes to ECP confluence | ||
** If you have not set up multi-factor authentication (MFA), you need to do that | ** If you have not set up multi-factor authentication (MFA), you need to do that | ||
Line 45: | Line 46: | ||
== Roundtable == | == Roundtable == | ||
+ | |||
+ | Matt has brought up some issues with compiling Ascent (more specifically, VTK-h) with newer versions of CMake. It is not totally clear what versions of CMake the issue comes up, but it has been noted that the dashboards do not use the latest versions. In particular, the highest version of CMake used for a CUDA build is 3.16. We need to check for newer versions of CMake. | ||
Vicente will work on making a release. Matt has requested the release happen right before the removal of virtuals from implicit functions. We will start with a branch and a release candidate that Matt can add fixes to. Matt will be something of a guinea pig for this. | Vicente will work on making a release. Matt has requested the release happen right before the removal of virtuals from implicit functions. We will start with a branch and a release candidate that Matt can add fixes to. Matt will be something of a guinea pig for this. |
Latest revision as of 13:01, 24 March 2021
Attendees: Ken Moreland (SNL), Nick Davis (SNL), Ollie Lo (LANL), Abhishek Yenpure (UO), Hank Childs (UO), Berk Geveci (Kitware), Sujin Philip (Kitware), Silvio Rizzi (ANL), David Pugmire (ORNL), Matt Larsen (LLNL), Vicente Bolea (Kitware), Terry Turton (LANL)
ECP Updates
- Upcoming OLCF early access machine: Spock
- No NDA
- We will automatically get access through our Summit allocation
- Expected to be available starting around mid May
- New ALCF early access machine: Arcticus
- Available now
- If you have an existing JLSE account, you should now have one on Articus.
- Security changes to ECP confluence
- If you have not set up multi-factor authentication (MFA), you need to do that
- I personally use an external authenticator app rather than the special SecSign ID
- They will start to disable accounts due to inactivity (1 month)
- If you need confluence access but don't use it much (as I expect many of you do), keep an eye on your email
- If you have not set up multi-factor authentication (MFA), you need to do that
- ECP Annual Meeting
- BoF days will be held 2 weeks beforehand (March 30 - April 1)
- Sign up for any BoFs you are interested in. (You don't need to be a member of ECP.)
- Go to: https://www.exascaleproject.org/event/ecp-community-bof-days-2021/
- In particular, consider attending the DAV session on March 30.
- Registration is open
- Will be virtual, registration is free
- Everyone funded by ECP/VTK-m should register and plan to attend at least part of the meeting
- Our project will need to create a poster
- Looks like the design will be a traditional poster
- Not totally clear how the poster session will work
- Looks like they are planning on using a virtual meet space.
- Let's think about possible highlights
- App engagement
- Success with porting (e.g. run on Aurora)
- Engagement with other ST teams (e.g. using Kokkos)
- WDM
- Submitted highlight on integration with EFFIS
- Expect to submit another highlight when used in a big run
- Possible highlight slide on particle advection with WarpX (Matt).
Porting Activities
- https://confluence.exascaleproject.org/display/STDA05/Porting+Activities+A21
- https://confluence.exascaleproject.org/display/STDA05/Porting+Activities+O21
ECP task updates
Roundtable
Matt has brought up some issues with compiling Ascent (more specifically, VTK-h) with newer versions of CMake. It is not totally clear what versions of CMake the issue comes up, but it has been noted that the dashboards do not use the latest versions. In particular, the highest version of CMake used for a CUDA build is 3.16. We need to check for newer versions of CMake.
Vicente will work on making a release. Matt has requested the release happen right before the removal of virtuals from implicit functions. We will start with a branch and a release candidate that Matt can add fixes to. Matt will be something of a guinea pig for this.
Wiki will go away at some point (and will definitely move soon). We should avoid using it for anything other than these notes.
We need some redesign design workflow for branchy release process
- When we create a VTK-m release, we have a branch for that release (at least until the subsequent release).
- New feature development continues on the main branch
- _Bug fixes_ happen off the release branch
- Get merged to both the release branch and the master branch
- This in principal should already be supported by GitLab
- Just create a merge request on a different branch
- Will need to add support to gitlab-push
- Ask VTK folks. They might just have a select few to do a special process.
- Need to check that bot stuff works (CI builds, Do: merge, etc.)
Should think about releasing VTK-m 1.6, 2.0.
Filter redesign requirements: https://gitlab.kitware.com/vtk/vtk-m/-/issues/601
Matt working on WarpX highlight.
Rename master branch?
- Need to wait for some changes to git before the transition (for some automated features of repo).
- Can we for a time have a master that mirrors the new default?
- Don't know, but seems possible. Will have to look into that.