search

NOAA-GFDL / SIS2

NOAA-GFDL's Sea Ice Simulator version 2
Other
16 stars 43 forks source link

Fix for specified ice layout non repro issue #200

Closed nikizadehgfdl closed 1 year ago

nikizadehgfdl commented 1 year ago
MJHarrison-GFDL commented 1 year ago

Approved.

marshallward commented 1 year ago

There seem to be regressions in two of the tests:

coupled_AM2_LM3_SIS2/Concurrent_ice_1deg/ocean.stats.gnu: FAILED
coupled_AM2_LM3_SIS2/Intersperse_ice_1deg/ocean.stats.gnu: FAILED

~Oddly some compilers and mem layouts seem to pass and others fail.~ (Nevermind, seems pretty consistent over compilers. The odd ordering on the page confused me.)

If you have access, you can see which ones here: https://gitlab.gfdl.noaa.gov/ogrp/SIS2/-/pipelines/20620

uramirez8707 commented 1 year ago

Thanks @nikizadehgfdl

This fixed the issue for the AM5 amip experiment. It now reproduces across layouts.

uramirez8707 commented 1 year ago

@marshallward any idea when this is going to be merged?

marshallward commented 1 year ago

@uramirez8707 We still need to address the regression raised above. AFAIK it is still happening. I will do this test again when I get some time.

marshallward commented 1 year ago

I looked further into this. The Ice%sCS%specified_ice caused a segmentation fault in the two failing tests.

I am guessing that it was not properly allocated or associated. @nikizadehgfdl can you look into this and try to fix it?

marshallward commented 1 year ago

Maybe you can fix this with an if (associated(Ice%sCS)) then ... but I will leave it to you all to decide.

marshallward commented 1 year ago

@nikizadehgfdl I can merge this once the possibility of an unallocated Ice@sCS is handled. Do you still need this?

marshallward commented 1 year ago

Gaea regression: https://gitlab.gfdl.noaa.gov/ogrp/SIS2/-/pipelines/21183 :heavy_check_mark: :yellow_circle: