Segmentation Fault geogrid.exe

[gfarache]

Hello
I'm trying to get some output variables of geogrid process in high-resolution to South America, but the geogrid is raising Segmentation Fault error. I'm attaching here the namelist.wps file that I've made with WRF Domain Wizard. Is there any limitation in the number of cells in geogrid? I'm asking it because I known this is a mesh larger than usual.

Thank you in advance

Parsed 49 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 )

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x149d85c25842 in ???
#1 0x149d85c249d5 in ???
#2 0x149d8583ebef in ???
#3 0x149d85990b3c in ???
#4 0x463eb0 in ???
#5 0x48f379 in ???
#6 0x41f150 in ???
#7 0x434aa7 in ???
#8 0x405b64 in ???
#9 0x149d858295cf in ???
#10 0x149d8582967f in ???
#11 0x4047a4 in ???
#12 0xffffffffffffffff in ???

 

[Ming Chen]

This is a memory issue. Your case is extremely large with the grid number of 13284 x 18927. For such a large number of grids, you have to compile and run WPS in parallel mode.

 

[gfarache]

This is a memory issue. Your case is extremely large with the grid number of 13284 x 18927. For such a large number of grids, you have to compile and run WPS in parallel mode.

Thank you for the answer, Ming. I'm already running in parallel mode, with 128 MPI processes. Should I run with more processes to generate this mesh?
I tried with 256 MPI processes but the problem persists.

 

[Ming Chen]

How many processors do you have? For such a big case, I would suggest you run with as many as possible processors.

Please note that running geogrid.exe requires much less memory than running wrf.exe. Once you reach the stage to run wrf.exe, I expect memory again could be a challenging issue.