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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Hi Tom, > I've been back working on this since I received your last message. Your > success at writing the larger file from the Fortran program on bluesky > strongly indicated that I'd either installed things incorrectly or > there's something different about our system here. I reinstalled netCDF > using your precise environment variable specifications. The wrf_test.f > job aborted at the same 2.147Gb size. That almost had to mean it was > some sort of system problem. I sat down with one of our IBM > consultants. He could not determine why there should be such a limit in > writing out that file since we routinely write out much larger (binary) > files. We have a small IBM test system here that has somewhat different > specifications than the large operational ones. He moved everything > over to that test system and successfully produced the netCDF output > file with 2.66Gb that you produced. He is sending this information to > IBM to find out why the problem exists on our production machines. In case it helps isolate the error, I've simplified the configuration a little bit, to get something minimal that works for us on the large file tests: CC=/usr/vac/bin/xlc CFLAGS='-g -q64 -D_LARGE_FILES' FC=/bin/xlf FFLAGS='-g -q64' F90=/bin/xlf90 F90FLAGS='-g -q64 -qsuffix=f=f90' CXX=/usr/vacpp/bin/xlC CXXFLAGS='-g -q64 -D_LARGE_FILES' ARFLAGS='-X64 cru' NMFLAGS=-X64 --Russ