cpld_bmark_v16 regression test ice_in inconsistency

[MinsukJi-NOAA]

Description

When cpld_bmark_v16 regression test runs, the block_size_y seems incorrect.

    nprocs            = 48
    nx_global         = 1440
    ny_global         = 1080
    block_size_x      = 60
    block_size_y      = 160 

Shouldn't block_size_y be 540?

Nevertheless, the simulation completes successfully.

To Reproduce:

$ git clone --recurse-submodules https://github.com/ufs-community/ufs-weather-model
$ cd ufs-weather-model/tests
$ ./rt.sh -n cpld_bmark_v16 -k >out 2>&1 &

Go to $RUNDIR and take a look at lines 158 to 162 of ice_in

[DeniseWorthen]

Yes, of course. Good catch. I'm amazed it ran and reproduced. Previously using edit.inputs, only the nproc needed to be changed if the default was changed. When I switched to atparse, I need to also export a new np2 and then a new BLCKX and BLCKY values when these are changed from the defaults. Does that seem right?

[MinsukJi-NOAA]

I see that tests/tests/cpld_bmark_v16 has export_cpl. Also that ice_in_template has BLCKX and BLCKY: looks like they should work -- a little confused now.

[MinsukJi-NOAA]

FYI, block_size_y is correct here: /scratch1/NCEPDEV/stmp2/Denise.Worthen/FV3_RT/REST_V16

[DeniseWorthen]

Yes, I made that sandbox prior to the atparse commit I believe.

Since we're using slenderx2, the Y-dimension should be split in 2; the X-dimension gets 1/2 of the total number of procs. After the atparse commit, for test where we've change NPROC_ICE, the test needs to have these lines added:

export np2=`expr $NPROC_ICE / 2`
export BLCKX=`expr $NX_GLB / $np2`
export BLCKY=`expr $NY_GLB / 2`

[MinsukJi-NOAA]

Yes, I made that sandbox prior to the atparse commit I believe.

Since we're using slenderx2, the Y-dimension should be split in 2; the X-dimension gets 1/2 of the total number of procs. After the atparse commit, for test where we've change NPROC_ICE, the test needs to have these lines added:

export np2=`expr $NPROC_ICE / 2`
export BLCKX=`expr $NX_GLB / $np2`
export BLCKY=`expr $NY_GLB / 2`

That makes sense.

[MinsukJi-NOAA]

Will this require input data change?

[DeniseWorthen]

no