[Wrf-users] RE: Segmentation fault with WRF-NMM
Kemp, Eric M.
Eric.Kemp at ngc.com
Wed Feb 6 07:40:53 MST 2008
Update on WRF-NMM: I have added the extra urban variables (this was also suggested by Ligia Bernardet at NOAA/ESRL, thanks), and the program now calls the NOAH soil model. However, I am now encountering an array out of bounds error in the RRTM radiation package. I will send more information shortly.
BTW, it is not clear to me why adding the urban variables is necessary; subroutines SURFACE_DRIVER and
LSM both have these variables listed as optional. Is this a compiler bug?
-Eric
Eric M. Kemp
Meteorologist
Northrop Grumman Information Technology
Intelligence Group (TASC)
4801 Stonecroft Boulevard
Chantilly, VA 20151
(703) 633-8300 x7078 (lab)
(703) 633-8300 x8278 (office)
(703) 449-3400 (fax)
eric.kemp at ngc.com
-----Original Message-----
From: WRF Help [mailto:wrfhelp at ucar.edu]
Sent: Tue 2/5/2008 12:52 PM
To: Kemp, Eric M.
Subject: Re: Segmentation fault with WRF-NMM
Hi Eric,
Could you try the following. In
dyn_nmm/module_PHYSICS_CALLS.F declare as integers
NUM_ROOF_LAYERS, NUM_WALL_LAYERS and NUM_ROAD_LAYERS,
define them to be equal to grid%num_soil_layers just
before call to SURFACE_DRIVER, and add them to the
argument list.
For example:
!
LOGICAL :: E_BDY,WARM_RAIN
!
! urban related variables
INTEGER :: ucmcall
INTEGER :: NUM_ROOF_LAYERS, NUM_WALL_LAYERS,
NUM_ROAD_LAYERS ! urban
!
!---------------------------------------------------------------------
.....
!
!-----------------------------------------------------------------
! urban related variable are added to arguments of
surface_driver
!-----------------------------------------------------------------
NUM_ROOF_LAYERS = grid%num_soil_layers !urban
NUM_WALL_LAYERS = grid%num_soil_layers !urban
NUM_ROAD_LAYERS = grid%num_soil_layers !urban
!-----------------------------------------------------------------------
!
!*** CALL SURFACE LAYER AND LAND SURFACE PHYSICS
!
!-----------------------------------------------------------------------
!
CALL
SET_TILES(GRID,IDS,IDE-1,JDS+1,JDE-1,ITS,ITE,JTS,JTE)
!
CALL SURFACE_DRIVER(
&
&
ACSNOM=ACSNOM,ACSNOW=ACSNOW,AKHS=AKHS,AKMS=AKMS
&
&
,ALBEDO=ALBEDO,BR=BR,CANWAT=CMC,CHKLOWQ=CHKLOWQ
&
&
,DT=DT,DX=DX,DZ8W=DZ,DZS=DZSOIL,GLW=RLW_DN_SFC
&
&
,GRDFLX=GRNFLX,GSW=RSW_NET_SFC,SWDOWN=RSW_DN_SFC
&
& ,GZ1OZ0=GZ1OZ0,HFX=TWBS
&
&
,HT=SFCZ,IFSNOW=IDUMMY,ISFFLX=ISFFLX,ISLTYP=ISLTYP
&
&
,ITIMESTEP=NTSD,IVGTYP=IVGTYP,LOWLYR=LOWLYR
&
& ,MAVAIL=ONE,RMOL=RMOL,MOL=MOL
&
& ,NUM_SOIL_LAYERS=NSOIL,P8W=P8W
&
&
,PBLH=PBLH,PI_PHY=PI_PHY,PSHLTR=PSHLTR,PSIH=PSIH
&
&
,PSIM=PSIM,P_PHY=P_PHY,Q10=Q10,Q2=Q2X,QFX=QWBS,QSFC=QS
&
& ,QSHLTR=QSHLTR,QZ0=QZ0,RAINCV=RAIN
&
&
,RHO=RR,SFCEVP=SFCEVPX,SFCEXC=SFCEXC,SFCRUNOFF=SSROFF
&
&
,SMOIS=SMC,SMSTAV=SMSTAV,SMSTOT=SMSTOT,SNOALB=MXSNAL
&
&
,SNOW=SNOW,SNOWC=SNOWC,SNOWH=SNOWH,STEPBL=NPHS
&
& ,SST=SST,SST_UPDATE=SST_UPDATE
&
&
,TH10=TH10,TH2=TH2X,T2=T2,THZ0=THZ0,TH_PHY=TH_PHY
&
&
,TMN=TG,TSHLTR=TSHLTR,TSK=TSFC,TSLB=STC,T_PHY=T_PHY
&
&
,U10=U10,UDRUNOFF=BGROFF,UST=USTAR,UZ0=UZ0H
&
&
,U_FRAME=U_FRAME,U_PHY=U_PHY,V10=V10,VEGFRA=VGFRCK
&
& ,VZ0=VZ0H,V_FRAME=V_FRAME,V_PHY=V_PHY
&
&
,WARM_RAIN=WARM_RAIN,WSPD=WSPD,XICE=SICE,XICEM=SICE
&
&
,ISICE=GRID%LANDUSE_ISICE,ISWATER=GRID%ISWATER
&
&
,XLAND=XLAND,Z=Z,ZNT=Z0,ZS=SLDPTH,CT=CT,TKE_MYJ=TKE
&
&
,ALBBCK=ALBASE,LH=ELFLX,SH2O=SH2O,SHDMAX=SHDMAX
&
&
,SHDMIN=SHDMIN,Z0=Z0BASE,FLQC=FLQC,FLHC=FLHC
&
& ,PSFC=PSFC_OUT,EMISS=EPSR
&
&
,SF_SFCLAY_PHYSICS=CONFIG_FLAGS%SF_SFCLAY_PHYSICS
&
&
,SF_SURFACE_PHYSICS=CONFIG_FLAGS%SF_SURFACE_PHYSICS
&
& ,RA_LW_PHYSICS=CONFIG_FLAGS%RA_LW_PHYSICS
&
& ,UCMCALL=UCMCALL
&
&
,IDS=IDS,IDE=IDE,JDS=JDS,JDE=JDE,KDS=KDS,KDE=KDE
&
&
,IMS=IMS,IME=IME,JMS=JMS,JME=JME,KMS=KMS,KME=KME
&
& ,I_START=GRID%I_START,I_END=GRID%I_END
&
& ,J_START=GRID%J_START,J_END=GRID%J_END
&
& ,KTS=KTS,KTE=KTE,NUM_TILES=GRID%NUM_TILES
&
! Optional args
&
,QV_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QV),F_QV=F_QV
&
&
,QC_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QC),F_QC=F_QC
&
&
,QR_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QR),F_QR=F_QR
&
&
,QI_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QI),F_QI=F_QI
&
&
,QS_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QS),F_QS=F_QS
&
&
,QG_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QG),F_QG=F_QG
&
& ,RAINBL=RAINBL
&
! for RUCLSM
& ,QSG=QSG,QVG=QVG,QCG=QCG,SOILT1=SOILT1
&
&
,TSNAV=TSNAV,SMFR3D=SMFR3D,KEEPFR3DFLAG=KEEPFR3DFLAG
&
&
,POTEVP=POTEVP,SNOPCX=SNOPCX,SOILTB=SOILTB,SR=SR
&
! for URBAN
& ,NUM_ROOF_LAYERS=NUM_ROOF_LAYERS
& ! urban
& ,NUM_WALL_LAYERS=NUM_WALL_LAYERS
& ! urban
& ,NUM_ROAD_LAYERS=NUM_ROAD_LAYERS
& ! urban
Please let me know how this goes.
--wrfhelp
On Mon, 4 Feb 2008 15:15:38 -0500
"Kemp, Eric M." <Eric.Kemp at ngc.com> wrote:
>
> Dear WRFer's:
>
> I am trying to run WRF-NMM with RSL_LITE/MPI on an
>Itaninum-2 cluster. I have
> compiled the model (both version 2.2 and 2.2.1) with
>ifort 9.1.036 and icc
> 9.1.042. When I try running the model in real data
>mode, a segmentation
> fault occurs when calling the NOAH soil model.
>
> Here is the stack trace:
>
> forrtl: severe (174): SIGSEGV, segmentation fault
>occurred
> Image PC Routine
> Line Source
>
> wrf.exe.nmm 4000000004298211 module_sf_noahlsm
> 46 module_sf_noahlsm.f90
> wrf.exe.nmm 4000000004923730 module_surface_dr
> 692 module_surface_driver.f90
> wrf.exe.nmm 4000000004823110 module_physics_ca
> 1016 module_PHYSICS_CALLS.f90
> wrf.exe.nmm 40000000033307E0 solve_nmm_
> 2726 solve_nmm.f90
> wrf.exe.nmm 40000000027293C0 solve_interface_
> 606 solve_interface.f90
> wrf.exe.nmm 40000000023BD180 module_integrate_
> 313 module_integrate.f90
> wrf.exe.nmm 40000000009E1040 module_wrf_top_mp
> 265 module_wrf_top.f90
> wrf.exe.nmm 40000000000064F0 MAIN__
> 24 wrf.f90
> wrf.exe.nmm 4000000000006380 Unknown
> Unknown Unknown
> libc.so.6.1 20000000008A1430 Unknown
> Unknown Unknown
> wrf.exe.nmm 4000000000006100 Unknown
> Unknown Unknown
>
> I do not experience this problem when using the older
>NMM soil model (option
> 99 in the namelist file). Also, I have been successful
>in running NMM on an AMD Opteron
> cluster using the Portland Group compilers.
>
> Has anyone else experienced this problem?
>
> -Eric
>
> Eric M. Kemp
> Meteorologist
> Northrop Grumman Information Technology
> Intelligence Group (TASC)
> 4801 Stonecroft Boulevard
> Chantilly, VA 20151
> (703) 633-8300 x7078 (lab)
> (703) 633-8300 x8278 (office)
> (703) 449-3400 (fax)
> eric.kemp at ngc.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20080206/bed97483/attachment.html
More information about the Wrf-users
mailing list