Updates to `1deg_jra55do_ryf` #91

dougiesquire · 2024-07-29T00:22:10Z

This PR:

Updates to a dev build of access-om3 that uses FMS without the large_file and deprecated_io flags set, and configures MOM to use netcdf4 - see Netcdf format from MOM6 COSIMA/access-om3#188
Moves to using environment modules to specify the exe in config.yaml - see Use Payu's new ability to get executable from path loaded by environment module COSIMA/access-om3#168. Note the module name includes the access-om3 commit hash.
Renames the input directory from 'input' to 'INPUT' - see Next Payu release will break our configurations COSIMA/access-om3#170
Adds new Payu userscripts from om3-scripts - see Updates to config post-processing COSIMA/access-om3#182:
- Replace local archive/setup_cice_restarts.sh with those in om3-scripts
- Concatenate daily CICE output to monthly
- Auto-generate Intake-ESM datastores

I'll try cherry-pick to other configs once this is merged

anton-seaice · 2024-07-29T05:02:14Z

We should update the wiki, and OCN_modelio:: in nuopc.runconfig (Can we set all the lines starting with pio to unused?) (i.e. COSIMA/access-om3#185)

anton-seaice · 2024-07-29T05:42:27Z

I ran this as written:

ll -h archive/output000/access-om3.mom6.*
-rw-r----- 1 as2285 tm70 185M Jul 29 15:15 archive/output000/access-om3.mom6.h.native_1900_01.nc
-rw-r----- 1 as2285 tm70  80M Jul 29 15:15 archive/output000/access-om3.mom6.h.sfc_1900_01.nc
-rw-r----- 1 as2285 tm70 847K Jul 29 15:15 archive/output000/access-om3.mom6.h.static.nc
-rw-r----- 1 as2285 tm70  97M Jul 29 15:15 archive/output000/access-om3.mom6.h.z_1900_01.nc

with payu walltime of 1055s

and then ran with deflate=1:

-rw-r----- 1 as2285 tm70 211M Jul 29 15:35 archive/output001/access-om3.mom6.h.native_1900_02.nc
-rw-r----- 1 as2285 tm70  79M Jul 29 15:35 archive/output001/access-om3.mom6.h.sfc_1900_02.nc
-rw-r----- 1 as2285 tm70 993K Jul 29 15:35 archive/output001/access-om3.mom6.h.static.nc
-rw-r----- 1 as2285 tm70 111M Jul 29 15:35 archive/output001/access-om3.mom6.h.z_1900_02.nc

with payu walltime of 1022s

So it looks like the extra time for compression is maybe worth it. Noting tuning the IO layout might improve this.

anton-seaice · 2024-07-29T06:08:36Z

I ran this as written:

ll -h archive/output000/access-om3.mom6.*
-rw-r----- 1 as2285 tm70 185M Jul 29 15:15 archive/output000/access-om3.mom6.h.native_1900_01.nc
-rw-r----- 1 as2285 tm70  80M Jul 29 15:15 archive/output000/access-om3.mom6.h.sfc_1900_01.nc
-rw-r----- 1 as2285 tm70 847K Jul 29 15:15 archive/output000/access-om3.mom6.h.static.nc
-rw-r----- 1 as2285 tm70  97M Jul 29 15:15 archive/output000/access-om3.mom6.h.z_1900_01.nc

with payu walltime of 1055s

and then ran with deflate=1:

-rw-r----- 1 as2285 tm70 211M Jul 29 15:35 archive/output001/access-om3.mom6.h.native_1900_02.nc
-rw-r----- 1 as2285 tm70  79M Jul 29 15:35 archive/output001/access-om3.mom6.h.sfc_1900_02.nc
-rw-r----- 1 as2285 tm70 993K Jul 29 15:35 archive/output001/access-om3.mom6.h.static.nc
-rw-r----- 1 as2285 tm70 111M Jul 29 15:35 archive/output001/access-om3.mom6.h.z_1900_02.nc

with payu walltime of 1022s

So it looks like the extra time for compression is maybe worth it. Noting tuning the IO layout might improve this.

Oh february has 28 days and january has 31 ... the time difference might be unrelated to compression

dougiesquire

~~I don't think this will run will it?~~ Scratch that, I misread your change. I assume you tried to run with these changes?

anton-seaice · 2024-07-29T06:45:10Z

Yes it works (or im having a 5 oclock moment!). It only removes lines for components which don't exist

dougiesquire · 2024-07-29T06:47:02Z

Let me know when you're happy and I'll do another test run with your added changes

input.nml

anton-seaice · 2024-07-29T06:57:26Z

Nothing further @dougiesquire

dougiesquire · 2024-07-30T11:17:47Z

Thanks @anton-seaice. Test run ran successfully. But I'm going to bundle in a couple more changes to close COSIMA/access-om3#168 and COSIMA/access-om3#170 to reduce the amount of cherry-picking I have to do. I won't include changes for COSIMA/access-om3#194 in this PR since this requires more testing.

github-actions · 2024-07-30T11:18:01Z

Thanks @anton-seaice. Test run ran successfully. But I'm going to bundle in a couple more changes to close COSIMA/access-om3#168 and COSIMA/access-om3#170 to reduce the amount of cherry-picking I have to do. I won't include changes for COSIMA/access-om3#194 in this PR since this requires more testing.

Automatic cherry-pick failed. Incorrect cherry-pick command:

the command is missing the into separator.

dougiesquire · 2024-07-30T11:19:35Z

^ 👀 @micaeljtoliveira

micaeljtoliveira · 2024-07-30T22:52:39Z

Oh no... We need to change the parsing of the comment to make sure we only trigger the workflow if the expression "cherry-pick" is at the beginning of the line.

github-actions · 2024-07-30T22:53:00Z

Oh no... We need to change the parsing of the comment to make sure we only trigger the workflow if the expression cherry-pick is at the beginning of the line.

Automatic cherry-pick failed. Incorrect cherry-pick command:

the command is missing the into separator.

dougiesquire · 2024-07-30T22:57:19Z

Maybe worth having a character at the start to distinguish the command? E.g. Tommy's versioning commands start with a "!".

aekiss · 2024-07-31T23:22:26Z

also trigger only for !cherry-pick (with trailing space) or !cherry-pick into ?

micaeljtoliveira · 2024-08-01T00:38:21Z

I think just adding the exclamation mark is enough to avoid 99.99% of unwanted triggers of the command.

Triggering it when not specifying into is a good thing, as it gives some feedback to the user that might otherwise be left wondering why the comment didn't trigger the workflow.

micaeljtoliveira · 2024-08-01T00:44:34Z

See #92 for the fix.

micaeljtoliveira · 2024-08-01T01:21:39Z

It should now be safe to discuss cherry-picks in PRs without being interrupted by the github bot.

aekiss · 2024-08-01T02:16:08Z

cherry-pick cherry-pick cherry-pick cherry-pick cherry-pick cherry-pick cherry-pick cherry-pick

aekiss · 2024-08-01T02:16:34Z

seems to work :-)

dougiesquire · 2024-08-02T11:26:28Z

@anton-seaice, I've bundled a couple more things into this PR with the hope it might reduce the amount of cherry-picking and PRs. Let me know if you're happy and I'll do some history rewriting to simplify the cherry-picking.

anton-seaice

Can you update the wiki please for where to set MOM IO & netcdf configs ?
I suggest we put this in four commits, and link to the related issue for each PR in the commit / description.

Is there somewhere obvious to document it needs payu/dev or payu 1.1.4 to run ? (in config.yaml, in the README, somewhere else?)

I haven't tested, but expect its fine.

…cess-om3#168)

Flag OCN_modelio::pio_* params as unused in nuopc.runconfig remove extra lines in nuopc.runconfig update comment for ocean nc type

- work-around symlink restart bug in CICE - concatenate CICE daily output - auto-generate an intake-esm datastore

anton-seaice · 2024-08-07T04:32:46Z

This looks good:

ncdump of MOM history and restart files shows format = netcdf-4 , shuffle = true, chunksize exist and are full size of x/y/time etc, deflate is 4.
payu finds the exe using the module, and the module name contains the access-om3 git commit used for the build. full binary path is in the manifests
folder is named INPUT and no errors are given loading forcings / grids etc
setup user scripts copies cice restart into work folder but there no extra copies in the output archive, intake esm datastore exists. We have however overwritten the monthly cice output with the daily cice output !

anton-seaice · 2024-08-07T04:35:30Z

Ill update the cice config and add to this PR

anton-seaice · 2024-08-07T06:50:13Z

For reasons that are unclear, the hist_suffix namelist config in cice appears to only support one character per history frequency. I can set this to 'd' and 'm' to progress the PR, or we can park it. Ill raise an issue tomorrow.

anton-seaice · 2024-08-08T01:23:06Z

This is blocked on COSIMA/access-om3#201 for now

anton-seaice · 2024-08-08T07:04:13Z

I suggest we drop the userscripts from this PR and merge the rest :)

dougiesquire · 2024-08-09T00:08:15Z

We decided in a meeting today to set hist_suffix to:

'.' for daily files
'm' for monthly files

The CICE filenames will look a bit yucky until COSIMA/access-om3#201 is resolved, but we can then at least use the concatenation userscript without modification.

…er at this point, and the userscript to concat ice daily output expects the default name to be used. See COSIMA/access-om3#201

anton-seaice · 2024-08-09T00:22:19Z

* '.' for daily files

sadly i was wrong and it needs to be blank until we change the userscripts see COSIMA/om3-scripts#21

anton-seaice

I've done the old change and approve trick :). This works now.

dougiesquire · 2024-08-09T00:26:02Z

sadly i was wrong and it needs to be blank until we change the userscripts see COSIMA/om3-scripts#21

I vote we change the userscript now to something that can work with both:

the *1day.mean* format
some workaround format for now

Is that possible? Then we won't end up with more dependent changes down the track

anton-seaice · 2024-08-09T00:47:06Z

sadly i was wrong and it needs to be blank until we change the userscripts see COSIMA/om3-scripts#21

I vote we change the userscript now to something that can work with both:
* the `*1day.mean*` format

* some workaround format for now
Is that possible? Then we won't end up with more dependent changes down the track

Yep "should" be fine. We can do that change at any point going forward - it doesn't impact this PR. It just needs to be done before changing the filenames to the desired filenames.

dougiesquire · 2024-08-09T00:53:32Z

it doesn't impact this PR

True - I wasn't thinking clearly sorry

dougiesquire · 2024-08-09T01:01:06Z

!cherry-pick 34e6e16 a8bd3a1 b191c68 81262fa 615c049 into 1deg_jra55do_iaf

github-actions · 2024-08-09T01:01:19Z

!cherry-pick 34e6e16 a8bd3a1 b191c68 81262fa 615c049 into 1deg_jra55do_iaf

Automatic cherry-pick failed. Could not find commit(s) 34e6e16 a8bd3a1 b191c68 81262fa 615c049 in 1deg_jra55do_ryf.

dougiesquire · 2024-08-09T01:04:48Z

!cherry-pick d24a5d2 2ee5b1d e152ad4 9f551ec 7beab25 into 1deg_jra55do_iaf

github-actions · 2024-08-09T01:05:12Z

Automatic Git cherry-picking of commit(s) d24a5d2 2ee5b1d e152ad4 9f551ec 7beab25 into 1deg_jra55do_iaf failed. This usually happens when cherry-picking results in a conflic or an empty commit. To manually cherry-pick the commits and open a pull request, please follow these instructions:

Create new branch from target branch:

git checkout 1deg_jra55do_iaf
git pull
git checkout -b cherry_pick_from_pr91_into_1deg_jra55do_iaf

Cherry-pick commits:

git cherry-pick d24a5d2c8f057ef3e7571fcde3abd6f2db13f753 2ee5b1d9e2852dcb8ab7a220e15d95eea09eceb6 e152ad48f0c9b41d9709b691cec1641275b6333a 9f551ec66734f5cd30c88f9e7479af4fdd000530 7beab257bfecb6d27227875dced02016f271a9ac

Fix any conflicts and/or empty commits by following the instructions provided by Git.
Push the new branch:

git push --set-upstream origin cherry_pick_from_pr91_into_1deg_jra55do_iaf

Open a new pull request on github making sure the target branch is set to 1deg_jra55do_iaf.

aekiss · 2024-08-09T01:44:02Z

@dougiesquire are you hitting merge conflicts due to the change in MOM_input format in 1deg_jra55do_ryf?

Should I expedite COSIMA/access-om3#200?

dougiesquire · 2024-08-09T01:55:13Z

There are conflicts in the MOM_input. But there are conflicts in every other modified file also, so expediting COSIMA/access-om3#200 won't help much here.

dougiesquire requested a review from anton-seaice July 29, 2024 00:22

dougiesquire commented Jul 29, 2024

View reviewed changes

anton-seaice reviewed Jul 29, 2024

View reviewed changes

input.nml Show resolved Hide resolved

dougiesquire changed the title ~~Use netcdf4 with MOM6: 1deg_jra55do_ryf~~ Updates to 1deg_jra55do_ryf Aug 2, 2024

dougiesquire mentioned this pull request Aug 2, 2024

Update input directory name from input to INPUT COSIMA/om3-scripts#20

Merged

anton-seaice reviewed Aug 5, 2024

View reviewed changes

dougiesquire added 3 commits August 6, 2024 10:25

Update exe (COSIMA/access-om3#188), use environment module (COSIMA/ac…

34e6e16

…cess-om3#168)

Use netcdf4 for MOM6 (COSIMA/access-om3#188)

a8bd3a1

Flag OCN_modelio::pio_* params as unused in nuopc.runconfig remove extra lines in nuopc.runconfig update comment for ocean nc type

Use Payu userscripts from om3-scripts (COSIMA/access-om3#182) to:

b191c68

- work-around symlink restart bug in CICE - concatenate CICE daily output - auto-generate an intake-esm datastore

dougiesquire force-pushed the 1deg_jra55do_ryf.issue188 branch from 1cfd33b to 929a26a Compare August 6, 2024 00:35

dougiesquire requested a review from anton-seaice August 6, 2024 00:41

anton-seaice force-pushed the 1deg_jra55do_ryf.issue188 branch from 929a26a to 81262fa Compare August 7, 2024 03:35

Set hist_suffix for cice history output. This can only be one charact…

615c049

…er at this point, and the userscript to concat ice daily output expects the default name to be used. See COSIMA/access-om3#201

anton-seaice approved these changes Aug 9, 2024

View reviewed changes

dougiesquire merged commit 7beab25 into 1deg_jra55do_ryf Aug 9, 2024

anton-seaice deleted the 1deg_jra55do_ryf.issue188 branch August 15, 2024 22:44

This was referenced Aug 21, 2024

1deg jra55do iaf.cherrypicks #102

Merged

025deg jra55do ryf.cherrypicks #103

Merged

1deg jra55do ryf.cherrypicks ACCESS-NRI/access-om3-wav-configs#44

Merged

1deg jra55do iaf.cherrypicks ACCESS-NRI/access-om3-wav-configs#45

Merged

Updates to 1deg_jra55do_ryf #91

Updates to 1deg_jra55do_ryf #91

Conversation

dougiesquire commented Jul 29, 2024 • edited by anton-seaice Loading

anton-seaice commented Jul 29, 2024 • edited Loading

anton-seaice commented Jul 29, 2024 • edited Loading

anton-seaice commented Jul 29, 2024

dougiesquire left a comment • edited Loading

Choose a reason for hiding this comment

anton-seaice commented Jul 29, 2024

dougiesquire commented Jul 29, 2024

anton-seaice commented Jul 29, 2024

dougiesquire commented Jul 30, 2024

github-actions bot commented Jul 30, 2024

dougiesquire commented Jul 30, 2024

micaeljtoliveira commented Jul 30, 2024

github-actions bot commented Jul 30, 2024

dougiesquire commented Jul 30, 2024

aekiss commented Jul 31, 2024

micaeljtoliveira commented Aug 1, 2024

micaeljtoliveira commented Aug 1, 2024

micaeljtoliveira commented Aug 1, 2024

aekiss commented Aug 1, 2024

aekiss commented Aug 1, 2024

dougiesquire commented Aug 2, 2024

anton-seaice left a comment

Choose a reason for hiding this comment

anton-seaice commented Aug 7, 2024

anton-seaice commented Aug 7, 2024

anton-seaice commented Aug 7, 2024

anton-seaice commented Aug 8, 2024

anton-seaice commented Aug 8, 2024

dougiesquire commented Aug 9, 2024

anton-seaice commented Aug 9, 2024

anton-seaice left a comment

Choose a reason for hiding this comment

dougiesquire commented Aug 9, 2024

anton-seaice commented Aug 9, 2024

dougiesquire commented Aug 9, 2024

dougiesquire commented Aug 9, 2024

github-actions bot commented Aug 9, 2024

dougiesquire commented Aug 9, 2024

github-actions bot commented Aug 9, 2024

aekiss commented Aug 9, 2024

dougiesquire commented Aug 9, 2024

Updates to `1deg_jra55do_ryf` #91

Updates to `1deg_jra55do_ryf` #91

dougiesquire commented Jul 29, 2024 •

edited by anton-seaice

Loading

anton-seaice commented Jul 29, 2024 •

edited

Loading

anton-seaice commented Jul 29, 2024 •

edited

Loading

dougiesquire left a comment •

edited

Loading