Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

(RESOLVED) geogrid.exe Suspend in Processing GREENFRAC

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

Yang_7Ke

New member
Hello

When I execute the ./geogrid.exe in the terminal.
It seems that the program is no longer running as the terminal displays the information: "Processing GREENFRAC".
It's wired, geogrid just can't finished its work without reporting any error.
I think the reason is that the polar regions are chosen.
And geogrid can be successful completed by choosing some mid-latitudes.

Then how I fix this problem?

The results of executing ./geogrid.exe and the content of namelist.wps are given as follow:

#--------------------- The results of executing ./geogrid.exe ---------------------------------------
Parsed 28 entries in GEOGRID.TBL
Processing domain 1 of 1
Processing XLAT and XLONG
Processing MAPFAC
Processing F and E
Processing ROTANG
Processing LANDUSEF
Calculating landmask from LANDUSEF ( WATER = 17 21 )
Processing HGT_M
Processing SOILTEMP
Processing SOILCTOP
Processing SCT_DOM
Processing SOILCBOT
Processing SCB_DOM
Processing ALBEDO12M
Processing GREENFRAC

#----------------------------------- namelist.wps-----------------------------------------------------------
&share
wrf_core = 'ARW',
max_dom = 1,
start_date = '2018-08-08_00:00:00','2006-08-16_12:00:00',
end_date = '2018-08-08_12:00:00','2006-08-16_12:00:00',
interval_seconds = 21600
io_form_geogrid = 2,
/
&geogrid
parent_id = 1, 1,
parent_grid_ratio = 1, 3,
i_parent_start = 1, 31,
j_parent_start = 1, 17,
e_we = 74, 112,
e_sn = 61, 97,
geog_data_res = 'default','default',
dx = 30000,
dy = 30000,
map_proj = 'polar',
ref_lat = -90,
ref_lon = 0,
truelat1 = -80,!
truelat2 = 60.0,!not work for polar
stand_lon = 0,
geog_data_path = '/home/yang7kewu/Public/WRF/Build_WRF/WPS_GEOG/'
/
 
The geogrid run for half a hour(maybe), then show me "Successful completion of geogrid."
How could it take so long time?
 
The GREENFRAC field is a "masked" field, in which we can interpolate only from land points in the input dataset and we interpolate only to land points in the model domain. When the land-sea mask in the GREENFRAC dataset doesn't agree with the LANDMASK field computed by the geogrid program, expensive searches through the GREENFRAC dataset are needed to assign valid values to land cells in your model domain.
 
What' more, would it be to help me save time by executing the following statements:
mpirun -np 8 ./geogrid.exe
 
I think you're correct that the lack of data (in this case, MODIS FPAR) over the Antarctic continent is leading to long processing times in the geogrid program. If the WPS were compiled with a "dmpar" option, you may see some improvement in processing time using more than one MPI task. Exactly how much of a speedup you'll see, I couldn't say; in any case, it's certainly worth a try.
 
I also had such a issue. Normally geogrid.exe finishes in a few minutes.
But for a nested domain covering part of Antarctic, geogrid.exe has been stuck at "Processing field 10 of 25 (GREENFRAC) " for 12 hours on a skx node.
see my domain configuration:
parent_grid_ratio = 1, 5, 3, 3, 3,
i_parent_start = 1, 93, 57, 66, 88,
j_parent_start = 1, 31, 50, 88, 231,
s_we = 1, 1, 1, 1, 1,
e_we = 305, 536, 256, 241, ,
s_sn = 1, 1, 1, 1, 1,
e_sn = 263, 716, 313, 343, 286,
geog_data_res = '5m','30s','30s',
dx = 15000,
dy = 15000,
map_proj = 'lambert',
ref_lat = -55.408,
ref_lon = 142.99,
ref_x = 152.5,
ref_y = 131.5,
truelat1 = -55.408,
truelat2 = -55.408,
stand_lon = 142.99,
Anyway to improve the WRF code?

Thanks
Xiao-MIng Hu
 
yuanfangcan said:
I also had such a issue. Normally geogrid.exe finishes in a few minutes.
But for a nested domain covering part of Antarctic, geogrid.exe has been stuck at "Processing field 10 of 25 (GREENFRAC) " for 12 hours on a skx node.
see my domain configuration:
parent_grid_ratio = 1, 5, 3, 3, 3,
i_parent_start = 1, 93, 57, 66, 88,
j_parent_start = 1, 31, 50, 88, 231,
s_we = 1, 1, 1, 1, 1,
e_we = 305, 536, 256, 241, ,
s_sn = 1, 1, 1, 1, 1,
e_sn = 263, 716, 313, 343, 286,
geog_data_res = '5m','30s','30s',
dx = 15000,
dy = 15000,
map_proj = 'lambert',
ref_lat = -55.408,
ref_lon = 142.99,
ref_x = 152.5,
ref_y = 131.5,
truelat1 = -55.408,
truelat2 = -55.408,
stand_lon = 142.99,
Anyway to improve the WRF code?

Thanks
Xiao-MIng Hu

Are you running geogrid with more than one MPI task (after having configured and compiled with a "dmpar" option), or are you running geogrid serially?
 
I compiled WPS with "Settings for Linux x86_64, Intel compiler (dmpar)"
but run WPS with
"
time geogrid.exe >& geogrid.log
...
ungrib.exe >& ungrib_presslevel.log
...
time ibrun metgrid.exe >& metgrid.log
"
I have been running WPS this way for a long time, I did not know geogrid.exe can be run with ibrun (multiple processors), which may be partially because it is not need previously because serial geogrid.exe normally finish in minutes/seconds.

Xiao-MIng Hu
 
It may be worth trying to run geogrid in parallel to see whether you get any useful speed-up in processing the GREENFRAC field.
 
mgduda said:
yuanfangcan said:
I also had such a issue. Normally geogrid.exe finishes in a few minutes.
But for a nested domain covering part of Antarctic, geogrid.exe has been stuck at "Processing field 10 of 25 (GREENFRAC) " for 12 hours on a skx node.
see my domain configuration:
parent_grid_ratio = 1, 5, 3, 3, 3,
i_parent_start = 1, 93, 57, 66, 88,
j_parent_start = 1, 31, 50, 88, 231,
s_we = 1, 1, 1, 1, 1,
e_we = 305, 536, 256, 241, ,
s_sn = 1, 1, 1, 1, 1,
e_sn = 263, 716, 313, 343, 286,
geog_data_res = '5m','30s','30s',
dx = 15000,
dy = 15000,
map_proj = 'lambert',
ref_lat = -55.408,
ref_lon = 142.99,
ref_x = 152.5,
ref_y = 131.5,
truelat1 = -55.408,
truelat2 = -55.408,
stand_lon = 142.99,
Anyway to improve the WRF code?

Thanks
Xiao-MIng Hu

Are you running geogrid with more than one MPI task (after having configured and compiled with a "dmpar" option), or are you running geogrid serially?

As for my case, using MPI seems help little.

That geogrid employs static data; in order to save computation time, you may recycle the same result (geo_em.*.nc) once generated by geogrid.
 
Top