This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
>To: address@hidden >From: Juliana Rew <address@hidden> >Subject: netCDF question (fwd) >Organization: UCAR/Unidata >Keywords: 200009072135.e87LZNN11083 Hi, > We received the following question from an SCD user. I verified that > his script runs on our Compaq platform, but when I try to run it > on our SGI or IBM SP (blackforest), it doesn't build a netcdf file at all > (much less a corrupted one, as he is reporting). Can you advise on how > we can debug this? Thanks, I suspect you would have to adjust the location of the library directory and maybe also the include directory and name of the Fortran compiler in the following line of the script: xlf -o test -I/usr/local/include test.f -L/usr/local/lib32/r4i4/ -lnetcdf to get it to run on the SGI or IBM SP. We don't have NCL installed here so we can't reproduce the error exactly. But Steve has a login on blackforest, so I'll forward this to him. --Russ From address@hidden Wed Sep 6 15:39:21 2000 From: "Marat Khairoutdinov" <address@hidden> Date: Wed, 6 Sep 2000 15:39:16 -0600 To: address@hidden, address@hidden Subject: netcdf problem on blackforest Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Length: 2701 Hi, I have the following problem with netcdf fortran calls on blackforest. I was able to strip it down to a simple fortran code generated by the script attached below. Some of the data get corrupted while written in netcdf file. I tested the same code on SGI and prospect (Compaq) - everything is just fine. Please could you pass this code to somebody to check what's going on. Thanks. -Marat ------------------------------------- Run the script below on blackforest. Note that although all elements of variables are set to zero, the Q2 variable contains garbage. #! /usr/bin/csh set echo cat >! test.f << EOF implicit none include 'netcdf.inc' character name*10 integer nzm, ntime, hbuf_length real f(2000) integer i,k integer vdimids(2), ndimids integer ncid,err,zid,timeid,varid f=0. nzm=64 ntime=31 err = NF_CREATE('tmp.nc', NF_CLOBBER, ncid) err = NF_REDEF(ncid) err = NF_DEF_DIM(ncid, 'time', ntime, timeid) err = NF_DEF_DIM(ncid, 'z', nzm, zid) err = NF_ENDDEF(ncid) ndimids=2 vdimids(1) = zid vdimids(2) = timeid do while(.true.) read(5,fmt=*,err=1000,end=1000) name write(6,'(a10)') name err = NF_REDEF(ncid) err = NF_DEF_VAR(ncid,name,NF_FLOAT, & ndimids,vdimids,varid) err = NF_ENDDEF(ncid) err = NF_PUT_VAR_REAL(ncid, varid, f) end do 1000 continue err = NF_CLOSE(ncid) end EOF cat >! input.txt << EOM PRES U V UOBS VOBS RHO MSE DSE SSE THETAE THETA THETAV THETAL TABS TABSOBS TL QT QV S QVOBS TTEND QTEND QC QI QR QS QG QN QP QSAT QCOND PRECIP RELH TLFLUX TLFLUXS TVFLUX QTFLUX QTFLUXS QCFLUX QIFLUX QPFLUX QPFLUXS SW SWSB UW UWSB VW VWSB RADLWUP RADLWDN RADSWUP RADSWDN RADQRLW RADQRSW RADQR RADQRS RADQRC Q1C Q2 U2 V2 W2 EOM # compile and run: xlf -o test -I/usr/local/include test.f -L/usr/local/lib32/r4i4/ -lnetcdf test < input.txt #----------------------------------------------------- # read netcdf file: cat >! read.ncl << EOF f=addfile("./tmp.nc","r") names = getfilevarnames(f) do i=0,dimsizes(names)-1 print(names(i)+" sum of all elements: "+sum(f->\$names(i)\$)) end do EOF ncl < read.ncl -- \\|||// | ^ ^ | ( o o ) _______oOOo______-______oOOo__ Marat F. Khairoutdinov, Ph.D. Department of Atmospheric Science Colorado State University Fort Collins, CO 80523, U.S.A. phone: (970) 491-8591 fax: (970) 491-8693 or 491-8428 address@hidden http://kiwi.atmos.colostate.edu/group/marat