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

[netCDF #XZZ-151006]: static/shared mode of compiling netcdf with intel and pgi compiler

Davide,

Sorry to take so long to respond to this important question.

> I believe this is a known issue, but it is going to become very serious.
> In bullets:
> 
> 1) I have been unable to compile the shared version of netcdf-4.1.3 (and
> I believe older versions, but it doesn't matter) on Lynx, using the
> intel and pgi compilers. I only succeeded with the --disable-shared flag
> - not big deal, I thought I was due to the "preference" of static
> environment by Cray
> 
> 2) investigating further, I discovered that others reported the very
> same problem on very different architecture, including plain-vanilla
> linux (e.g.
> http://www.unidata.ucar.edu/mailing_lists/archives/netcdfgroup/2011/msg00361.html
> ) - not big deal (yet)
> 
> 3) IBM won the procurement for NWSC machine, now called Yellowstone.
> Intel and PGI will be the production compilers on that machine (now the
> deal starts to be not so little)
> 
> 4) IBM will *NOT* allow static linking on that machine. We do have to
> build everything in shared mode (now, that's a very serious issue)
> 
> Do you have any insight?

Have you tried compiling with version 4.2-rc1 or with the current snapshot 
release?

Have you asked the vendors of the Intel or PGI compilers about the problem?

Another possibility is posing the question on the netcdfgroup mailing list.
If you decided to do that, it would be fine to say that Unidata has been
unable to solve the problem.

Sorry we can't be of more help at this time.  Maybe when I get back from an
ESRI meeting next week, I'll be able to look at this more closely ...

--Russ

Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: XZZ-151006
Department: Support netCDF
Priority: Normal
Status: Closed