Due to the current gap in continued funding from the U.S. National Science Foundation (NSF), the NSF Unidata Program Center has temporarily paused most operations. See NSF Unidata Pause in Most Operations for details.
Dan Kokron <daniel.kokron@xxxxxxxx> writes: > Hello all, > > I have been getting bad values in NetCDF-4 files created with > rc = nf_create_par("test.nc4", IOR(NF_CLOBBER,NF_NETCDF4), MPI_COMM_WORLD, > info, UNIT) > . > . > rc = NF_PUT_VARA(UNIT, dstid, start, cnt, f90ptr3) > > The contents of f90ptr3 on each process are okay yet many locations in > the file have values like 9.96920996839e+36 > > I get the odd values with netcdf-4.1.2-rc1 + hdf5-1.8.6 and go away with > netcdf-4.1.1 + hdf5-1.8.5-patch1 using the same application. > > I'd like to continue to narrow down the variables. Is it safe to use > hdf5-1.8.6 with netcdf-4.1.1? > > hdf5-1.8.5-patch1 with netcdf-4.1.2-rc1? Can you try upgrading to the 4.1.2 relaase? Or perhaps the 4.1.3-beta1 release? A sample program that shows the problem would be helpful... Thanks, Ed -- Ed Hartnett -- ed@xxxxxxxxxxxxxxxx
netcdfgroup
archives: