issue with R8-> R4 conversion

[Jim Edwards]

I am using the latest parallel-netcdf (svn revision 3824) and running into
a problem on an IBM pwr9 system with ibm xl compilers.
The issue is that writing a double-precision in memory variable to a single
precision file field results in NaNF in the file.   I tried the same
case using serial netcdf and the file was written correctly.

Any ideas?


-- 
Jim Edwards

CESM Software Engineer
National Center for Atmospheric Research
Boulder, CO
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/parallel-netcdf/attachments/20181018/9da1b11b/attachment.html>