Cannot place all ranks on node list

[RuiyuSun]

I am trying to conduct a high resolution (C1152) coupled experiment on dogwood. Because the forecast length is fairly long the default number of nodes (tasks) can't complete the gfsfcst job. The default layouts are
layout_gfs_x=8
layout_gfs_y=16.

I tried to increase the number of nodes by increasing the layouts. An example is the following pair.
layout_gfs_x=12
layout_gfs_y=24

But I had the following error.
Cannot place all ranks on node list

mpiexec -l -n 10240 -ppn 128 --cpu-bind depth --depth 1 /lfs/h2/emc/stmp/ruiyu.sun/RUNDIRS/HR43/gfs.2020012600/gfsfcst.2020012600/fcst.28576/ufs_model.

Any suggestion?

Thanks,
Ruiyu

[jiandewang]

@RuiyuSun
I cloned latset g-w and tried C1152 S2SW, it works fine for 24x16 and 12x24 layout. See my log file on dogwood
12x24: /lfs/h2/emc/ptmp/jiande.wang/HR4-work/HR4-20191212/COMROOT/2019121200/HR4-20191212/logs/2019121200
24x16: /lfs/h2/emc/ptmp/jiande.wang/HR4-work/HR4-20191203/COMROOT/2019120300/HR4-20191203/logs/2019120300

[jiandewang]

the g-w I cloned:
commit 49f697a (HEAD -> develop, origin/develop, origin/HEAD)
Author: DavidNew-NOAA 134300700+DavidNew-NOAA@users.noreply.github.com
Date: Sat Sep 7

[jiandewang]

my g-w
/lfs/h2/emc/couple/noscrub/Jiande.Wang/HR4-test/global-workflow

[RuiyuSun]

Thanks! @jiandewang I copied your experiment directory. It worked fine for 12X24 layout.

[RuiyuSun]

I also used your global workflow experiment.