[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

960130: netCDF on IEEE Cray T90

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.

  • Subject: 960130: netCDF on IEEE Cray T90
  • Date: Tue, 30 Jan 1996 10:41:32 -0700 (MST)
  • Disclaimer: I speak for NCAR/SCD, NOT!
> 
> Hi again-
> 
> This spring, GFDL will be getting an IEEE Cray T90, and I was curious
> as to whether we can expect netCDF 2.4 to build and work properly.  In
> particular, I was wondering if all the (*nice*) work done by Jeff Kuehn
> to optimize Cray performance will apply to an IEEE Cray.
> 
> Thanks in advance-
> John P. Sheldon
> 

Most of it will, but small changes may need to be made in the XDR
routines that convert the native floating point and integer format
to 32 bit IEEE.  The current version uses CRAY2IEG() and IEG2CRAY
for these conversions, and I suspect that there will need to be a
few #ifdef's added for the IEEE T90 so that it does a 64 bit IEEE
to 32 bit IEEE conversion instead of the current CRAY to 32 bit
IEEE conversion.  The changes in the XDR routines should be the 
only changes required.  When your T90 arrives, contact me.  We may
be able to work something out by which I'll code the necessary
changes.


--Jeff Kuehn, Group Head, NCAR/SCD Technical Consulting