<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7652.24">
<TITLE>RE: Segmentation fault with WRF-NMM</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<BR>
<P><FONT SIZE=2>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.<BR>
<BR>
BTW, it is not clear to me why adding the urban variables is necessary; subroutines SURFACE_DRIVER and<BR>
LSM both have these variables listed as optional. Is this a compiler bug?<BR>
<BR>
-Eric<BR>
<BR>
Eric M. Kemp<BR>
Meteorologist<BR>
Northrop Grumman Information Technology<BR>
Intelligence Group (TASC)<BR>
4801 Stonecroft Boulevard<BR>
Chantilly, VA 20151<BR>
(703) 633-8300 x7078 (lab)<BR>
(703) 633-8300 x8278 (office)<BR>
(703) 449-3400 (fax)<BR>
eric.kemp@ngc.com<BR>
<BR>
<BR>
<BR>
-----Original Message-----<BR>
From: WRF Help [<A HREF="mailto:wrfhelp@ucar.edu">mailto:wrfhelp@ucar.edu</A>]<BR>
Sent: Tue 2/5/2008 12:52 PM<BR>
To: Kemp, Eric M.<BR>
Subject: Re: Segmentation fault with WRF-NMM<BR>
<BR>
Hi Eric,<BR>
<BR>
Could you try the following. In<BR>
dyn_nmm/module_PHYSICS_CALLS.F declare as integers<BR>
NUM_ROOF_LAYERS, NUM_WALL_LAYERS and NUM_ROAD_LAYERS,<BR>
define them to be equal to grid%num_soil_layers just<BR>
before call to SURFACE_DRIVER, and add them to the<BR>
argument list.<BR>
<BR>
For example:<BR>
<BR>
!<BR>
LOGICAL :: E_BDY,WARM_RAIN<BR>
!<BR>
! urban related variables<BR>
INTEGER :: ucmcall<BR>
INTEGER :: NUM_ROOF_LAYERS, NUM_WALL_LAYERS,<BR>
NUM_ROAD_LAYERS ! urban<BR>
!<BR>
!---------------------------------------------------------------------<BR>
<BR>
.....<BR>
<BR>
<BR>
!<BR>
!-----------------------------------------------------------------<BR>
! urban related variable are added to arguments of<BR>
surface_driver<BR>
!-----------------------------------------------------------------<BR>
NUM_ROOF_LAYERS = grid%num_soil_layers !urban<BR>
NUM_WALL_LAYERS = grid%num_soil_layers !urban<BR>
NUM_ROAD_LAYERS = grid%num_soil_layers !urban<BR>
<BR>
!-----------------------------------------------------------------------<BR>
!<BR>
!*** CALL SURFACE LAYER AND LAND SURFACE PHYSICS<BR>
!<BR>
!-----------------------------------------------------------------------<BR>
!<BR>
CALL<BR>
SET_TILES(GRID,IDS,IDE-1,JDS+1,JDE-1,ITS,ITE,JTS,JTE)<BR>
!<BR>
CALL SURFACE_DRIVER(<BR>
&<BR>
&<BR>
ACSNOM=ACSNOM,ACSNOW=ACSNOW,AKHS=AKHS,AKMS=AKMS<BR>
&<BR>
&<BR>
,ALBEDO=ALBEDO,BR=BR,CANWAT=CMC,CHKLOWQ=CHKLOWQ<BR>
&<BR>
&<BR>
,DT=DT,DX=DX,DZ8W=DZ,DZS=DZSOIL,GLW=RLW_DN_SFC<BR>
&<BR>
&<BR>
,GRDFLX=GRNFLX,GSW=RSW_NET_SFC,SWDOWN=RSW_DN_SFC<BR>
&<BR>
& ,GZ1OZ0=GZ1OZ0,HFX=TWBS<BR>
&<BR>
&<BR>
,HT=SFCZ,IFSNOW=IDUMMY,ISFFLX=ISFFLX,ISLTYP=ISLTYP<BR>
&<BR>
&<BR>
,ITIMESTEP=NTSD,IVGTYP=IVGTYP,LOWLYR=LOWLYR<BR>
&<BR>
& ,MAVAIL=ONE,RMOL=RMOL,MOL=MOL<BR>
&<BR>
& ,NUM_SOIL_LAYERS=NSOIL,P8W=P8W<BR>
&<BR>
&<BR>
,PBLH=PBLH,PI_PHY=PI_PHY,PSHLTR=PSHLTR,PSIH=PSIH<BR>
&<BR>
&<BR>
,PSIM=PSIM,P_PHY=P_PHY,Q10=Q10,Q2=Q2X,QFX=QWBS,QSFC=QS<BR>
&<BR>
& ,QSHLTR=QSHLTR,QZ0=QZ0,RAINCV=RAIN<BR>
&<BR>
&<BR>
,RHO=RR,SFCEVP=SFCEVPX,SFCEXC=SFCEXC,SFCRUNOFF=SSROFF<BR>
&<BR>
&<BR>
,SMOIS=SMC,SMSTAV=SMSTAV,SMSTOT=SMSTOT,SNOALB=MXSNAL<BR>
&<BR>
&<BR>
,SNOW=SNOW,SNOWC=SNOWC,SNOWH=SNOWH,STEPBL=NPHS<BR>
&<BR>
& ,SST=SST,SST_UPDATE=SST_UPDATE<BR>
&<BR>
&<BR>
,TH10=TH10,TH2=TH2X,T2=T2,THZ0=THZ0,TH_PHY=TH_PHY<BR>
&<BR>
&<BR>
,TMN=TG,TSHLTR=TSHLTR,TSK=TSFC,TSLB=STC,T_PHY=T_PHY<BR>
&<BR>
&<BR>
,U10=U10,UDRUNOFF=BGROFF,UST=USTAR,UZ0=UZ0H<BR>
&<BR>
&<BR>
,U_FRAME=U_FRAME,U_PHY=U_PHY,V10=V10,VEGFRA=VGFRCK<BR>
&<BR>
& ,VZ0=VZ0H,V_FRAME=V_FRAME,V_PHY=V_PHY<BR>
&<BR>
&<BR>
,WARM_RAIN=WARM_RAIN,WSPD=WSPD,XICE=SICE,XICEM=SICE<BR>
&<BR>
&<BR>
,ISICE=GRID%LANDUSE_ISICE,ISWATER=GRID%ISWATER<BR>
&<BR>
&<BR>
,XLAND=XLAND,Z=Z,ZNT=Z0,ZS=SLDPTH,CT=CT,TKE_MYJ=TKE<BR>
&<BR>
&<BR>
,ALBBCK=ALBASE,LH=ELFLX,SH2O=SH2O,SHDMAX=SHDMAX<BR>
&<BR>
&<BR>
,SHDMIN=SHDMIN,Z0=Z0BASE,FLQC=FLQC,FLHC=FLHC<BR>
&<BR>
& ,PSFC=PSFC_OUT,EMISS=EPSR<BR>
&<BR>
&<BR>
,SF_SFCLAY_PHYSICS=CONFIG_FLAGS%SF_SFCLAY_PHYSICS<BR>
&<BR>
&<BR>
,SF_SURFACE_PHYSICS=CONFIG_FLAGS%SF_SURFACE_PHYSICS<BR>
&<BR>
& ,RA_LW_PHYSICS=CONFIG_FLAGS%RA_LW_PHYSICS<BR>
&<BR>
& ,UCMCALL=UCMCALL<BR>
&<BR>
&<BR>
,IDS=IDS,IDE=IDE,JDS=JDS,JDE=JDE,KDS=KDS,KDE=KDE<BR>
&<BR>
&<BR>
,IMS=IMS,IME=IME,JMS=JMS,JME=JME,KMS=KMS,KME=KME<BR>
&<BR>
& ,I_START=GRID%I_START,I_END=GRID%I_END<BR>
&<BR>
& ,J_START=GRID%J_START,J_END=GRID%J_END<BR>
&<BR>
& ,KTS=KTS,KTE=KTE,NUM_TILES=GRID%NUM_TILES<BR>
&<BR>
! Optional args<BR>
&<BR>
,QV_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QV),F_QV=F_QV<BR>
&<BR>
&<BR>
,QC_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QC),F_QC=F_QC<BR>
&<BR>
&<BR>
,QR_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QR),F_QR=F_QR<BR>
&<BR>
&<BR>
,QI_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QI),F_QI=F_QI<BR>
&<BR>
&<BR>
,QS_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QS),F_QS=F_QS<BR>
&<BR>
&<BR>
,QG_CURR=MOIST_TRANS(IMS,KMS,JMS,P_QG),F_QG=F_QG<BR>
&<BR>
& ,RAINBL=RAINBL<BR>
&<BR>
! for RUCLSM<BR>
& ,QSG=QSG,QVG=QVG,QCG=QCG,SOILT1=SOILT1<BR>
&<BR>
&<BR>
,TSNAV=TSNAV,SMFR3D=SMFR3D,KEEPFR3DFLAG=KEEPFR3DFLAG<BR>
&<BR>
&<BR>
,POTEVP=POTEVP,SNOPCX=SNOPCX,SOILTB=SOILTB,SR=SR<BR>
&<BR>
! for URBAN<BR>
& ,NUM_ROOF_LAYERS=NUM_ROOF_LAYERS<BR>
& ! urban<BR>
& ,NUM_WALL_LAYERS=NUM_WALL_LAYERS<BR>
& ! urban<BR>
& ,NUM_ROAD_LAYERS=NUM_ROAD_LAYERS<BR>
& ! urban<BR>
<BR>
Please let me know how this goes.<BR>
--wrfhelp<BR>
<BR>
On Mon, 4 Feb 2008 15:15:38 -0500<BR>
"Kemp, Eric M." <Eric.Kemp@ngc.com> wrote:<BR>
><BR>
> Dear WRFer's:<BR>
><BR>
> I am trying to run WRF-NMM with RSL_LITE/MPI on an<BR>
>Itaninum-2 cluster. I have<BR>
> compiled the model (both version 2.2 and 2.2.1) with<BR>
>ifort 9.1.036 and icc<BR>
> 9.1.042. When I try running the model in real data<BR>
>mode, a segmentation<BR>
> fault occurs when calling the NOAH soil model.<BR>
><BR>
> Here is the stack trace:<BR>
><BR>
> forrtl: severe (174): SIGSEGV, segmentation fault<BR>
>occurred<BR>
> Image PC Routine<BR>
> Line Source<BR>
><BR>
> wrf.exe.nmm 4000000004298211 module_sf_noahlsm<BR>
> 46 module_sf_noahlsm.f90<BR>
> wrf.exe.nmm 4000000004923730 module_surface_dr<BR>
> 692 module_surface_driver.f90<BR>
> wrf.exe.nmm 4000000004823110 module_physics_ca<BR>
> 1016 module_PHYSICS_CALLS.f90<BR>
> wrf.exe.nmm 40000000033307E0 solve_nmm_<BR>
> 2726 solve_nmm.f90<BR>
> wrf.exe.nmm 40000000027293C0 solve_interface_<BR>
> 606 solve_interface.f90<BR>
> wrf.exe.nmm 40000000023BD180 module_integrate_<BR>
> 313 module_integrate.f90<BR>
> wrf.exe.nmm 40000000009E1040 module_wrf_top_mp<BR>
> 265 module_wrf_top.f90<BR>
> wrf.exe.nmm 40000000000064F0 MAIN__<BR>
> 24 wrf.f90<BR>
> wrf.exe.nmm 4000000000006380 Unknown<BR>
> Unknown Unknown<BR>
> libc.so.6.1 20000000008A1430 Unknown<BR>
> Unknown Unknown<BR>
> wrf.exe.nmm 4000000000006100 Unknown<BR>
> Unknown Unknown<BR>
><BR>
> I do not experience this problem when using the older<BR>
>NMM soil model (option<BR>
> 99 in the namelist file). Also, I have been successful<BR>
>in running NMM on an AMD Opteron<BR>
> cluster using the Portland Group compilers.<BR>
><BR>
> Has anyone else experienced this problem?<BR>
><BR>
> -Eric<BR>
><BR>
> Eric M. Kemp<BR>
> Meteorologist<BR>
> Northrop Grumman Information Technology<BR>
> Intelligence Group (TASC)<BR>
> 4801 Stonecroft Boulevard<BR>
> Chantilly, VA 20151<BR>
> (703) 633-8300 x7078 (lab)<BR>
> (703) 633-8300 x8278 (office)<BR>
> (703) 449-3400 (fax)<BR>
> eric.kemp@ngc.com<BR>
><BR>
<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>