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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Chris, > To: address@hidden > From: "Chris Dallimore" <address@hidden> > Subject: netCDF Fortran (77 + 90) - Mac OSX 10.3 and Linux redhat - Strang 2d > array error > Organization: UWA > Keywords: 200407230232.i6N2WMv1022403 netCDF F90 MacOS-X The above message contained the following: > Institution: Centre for Water Research > Package Version: 3.5.1 > Operating System: Mac OSX 10.3 and Linux redhat > Hardware Information: g5 dual 2ghz; pentium 4 2.8ghz > Inquiry: I'm getting a strange error using the f90 Netcdf library on > two different OS using two different compilers. > > I am ouputting a two-dimensional real array to a netcdf file using > NF90_PUT_VAR. The array looks like > > 1 3 NAN > 2 4 5 > NAN NAN 6 > > Firstly on OSX, I have managed to configure make and test the library > using both the beta version of IBM's xlf and the release version of xlf > (8.1). > > Now comes the strang bit. When our program is linked to the beta > compiled library the array is output fine but when I link it to th > elibrary compiled with release version of the compiler the output is > > 1 NAN 4 > 2 NAN NAN > NAN 3 NAN > > And to complicate things further on Linux everything is OK if I use the > Intel compiler version 8 of lahey version 6.1 but I get the same > skewed result if I use Intel version 7. Very strange, indeed. Did all versions of the netCDF package pass the package's "make test" test? Is this a simple program that you could send us? > I am assuming this is probabley a compiler problem but wanted to let > you know and check to see if this sort of problem had been observed > before. We've seen array indexing problems with certain versions of the Portland Group Fortran-90 compiler (pgf90) but not (yet) with the xlf compiler. > thanks for your time. > > Chris Dallimore > > From: "Chris Dallimore" <address@hidden> > > Institution: CWR > Package Version: 3.5.1 > Operating System: OSX > Hardware Information: G5 > Inquiry: Hi, I sent an email earlier today about a problem on OSX. > > I made a mistake in my diagnosis. The problem with the xlf compiler > was not introduced between beta and release but between release and the > March 2004 update. The recent July 2004 update does not fix the > problem. > > Thanks > > Chris Dallimore Regards, Steve Emmerson > NOTE: All email exchanges with Unidata User Support are recorded in the > Unidata inquiry tracking system and then made publically available > through the web. If you do not want to have your interactions made > available in this way, you must let us know in each email you send to us.