Compiler warnings may not be suppressed #1984
Comments
|
I agree. There are groups that have done some of this work in their push from R2O, so maybe we can bring some of those edits back from operations. |
|
When I compile the code without or: If I disable just this warning 5462 (using CMEPS: stochastic_physics: ccpp/physics: atmos_cubed_sphere: MOM6: |
|
It's hard to believe that there are no warnings from either CICE or WW3? I wonder if these components are having the -nowarn overwritten. |
@DeniseWorthen looking into the cmake build scripts, I think you are likely correct WW3 is overriding this at the moment. |
I do not see where WW3 sets -nowarn. |
|
There are only 7 'remarks' in WW3, I'm not sure if NCO considers remarks as warnings. It would be nice to clean them as well (either by changing the code or disabling them) |
Yes, I could not find it either. WW3 is unsetting flags from parent components, so I was mostly confirming that a top level -nowarn would be overwritten. |
It is interesting there are only remarks, and no warnings. I will look into resolving those (though we have some bigger fish to fry, so not immediately). |
|
@DusanJovic-NOAA For CICE, the CMakeList shows So to me, this appears that CICE is not appending the fortran flags to any existing ones, but creating it's own list. Is that right? If I look at a compile line for CICE in the compile log, I don't see the -nowarn setting, which I should if it was appending to an existing list. I seem to remember this question before about each component setting it's own compiler settings vs a single set of compiler settings for the all the components. |
|
@jiandewang FYI. |
|
just communicated with GFDL, will start to eliminate them. |
|
@jiandewang Thanks! |
|
I recompiled the current develop branch and I still see more or less the same warnings (I didn't compare them line by line). But now I do see one warning in CICE: |
|
OK, thanks. I also have a fix for the one you found in CMEPS. How do you want to go about fixing these? I can make a feature branch which you could point to for testing? |
A feature branch should work. Thanks. |
|
@DusanJovic-NOAA Please try these feature branches: https://github.com/DeniseWorthen/CMEPS/commits/feature/nowarn/ Thanks |
I recompiled the code using these two branches and I do not see CICE and CMEPS warnings anymore. |
|
@DeniseWorthen would you please create PRs for CICE and CMEPS? Thanks |
|
@Qingfu-Liu May I ask if you can take a look at the warnings in CCPP physics? Thanks |
|
@junwang-noaa Yes, I'll create a PR for UWM to bring in these two updates. |
|
|
@jun Wang - NOAA Federal ***@***.***> The warnings listed in the
issue #1984 have the old CCPP physics directory structure. I will running
to duplicate the issues in the new CCPP physics directory structure
…On Mon, Feb 5, 2024 at 1:50 PM Denise Worthen ***@***.***> wrote:
@junwang-noaa <https://github.com/junwang-noaa> Yes, I'll create a PR for
UWM to bring in these two updates.
—
Reply to this email directly, view it on GitHub
<#1984 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGTS6UWAN26YPJXVAVKNUCLYSESWJAVCNFSM6AAAAAA7AIFCIOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMRXHAYDCNRZG4>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Just a note that, for the upcoming HAFSv2 upgrade, we are also required to address the build/compile warnings. Thanks for having already started working on these! |
|
@junwang-noaa I have created PR#169 to remove the compile warnings from CCPP physics in the ufs-community repository. |
|
Thanks, @Qingfu-Liu |
|
@DusanJovic-NOAA can you re-compile using the following MOM6 branch ? |
Thanks. I do not see MOM6 warnings anymore. |
|
@DusanJovic-NOAA thanks for your testing. Actually there is one warning pop out only on wcoss2 (not on HERA that I tested). Marshall will prvide an updated code and I will let you know when it's ready for a re-test. |
|
@DusanJovic-NOAA can you repeat your test but use latest commit (hash # 51fe0dab) on wcoss2 |
No warnings on wcoss2 with 51fe0dab |
|
@DusanJovic-NOAA thanks. I will signal GFDL to issue PR for MOM6 |
|
We encountered an intel compiler warning from fv3atm (on WCOSS2/Hera/Jet) in HAFS: The log information is not very helpful in locating the issue and unable to locate the "MY_POINTER" variable/function in the code. |
Currently in the top-level cmake and in atmos_cube_sphere cmake we use the
-nowarnflag (Intel compiler) which suppress all compiler warnings.According to "NCEP Central Operations WCOSS Implementation Standards" this is not allowed (page 15 section IV B. (5)):
"Code must compile without errors or warnings. Errors and warnings may not be suppressed, and the compiler warning level ("-W" options) must be at least the default one."
We should remove this flag (from ufs-weather-model and atmos_cubed_sphere) and start working on eliminating all warnings.
The text was updated successfully, but these errors were encountered: