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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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, >To: Carlie Coats <address@hidden> >cc: address@hidden, >cc: address@hidden >From: address@hidden (Mark Reed) >Subject: Re: who has experience with netCDF 3.3 on cray t3e ? >Organization: North Carolina Super Computing Center >Keywords: 199705271510.JAA14501 > > The problem on the T3D/T3E is exactly as you stated. The default floating > point size is 4 bytes (32 bits) for C and 8 bytes (64 bits) for Fortran. To > get around this problem, Atanas Trayanov and myself have written some very > simple ncvg wrapper routines to allow netcdf to work fairly seamlessly. In > addition, there is some preprocessing done on the files. If you are interested > in this, just contact Atanas (address@hidden) or myself (address@hidden) and > we will be glad to share this with you. I should point out that our work was > done for the T3D, before the f90 compiler was available. The f90 compiler will > allow either 32 or 64 bit reals. The default is 64 (unless you specify > real*4). Thanks for the offer. I'm hoping we will just fix cfortran.h to do the right thing with the default compilation. It seems like most CRAY fortran codes are going to want to use REAL and think that it is 64bit. I'm not a Cray expert particularly. I'm wondering whether you or someone there would be willing to work with Burkhard Burow <address@hidden> to get cfortran.h working on the t3e (and remain working on other Cray systems :-)). Since it's not my main gig and I have to log in remotely to a system in Maryland (gfdl) to test things, progress is slow. Thanks -glenn