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

Re: NetCDF 3.5.1 beta version on Cray T3E

Russ,

>Date: Sun, 22 Sep 2002 09:53:29 -0600
>From: Russ Rew <address@hidden>
>Organization: UCAR Unidata Program
>To: Unidata Support <address@hidden>
>Subject: Re: NetCDF 3.5.1 beta version on Cray T3E 

The above message contained the following:

> No, I'm sorry to say we never heard back on this issue so it's still
> unresolved.  And we still don't have access to a Cray T3E.  I'm CC:ing
> Steve Emmerson on this reply, in case he has any ideas for how to make
> progress.

I noticed that the line number of the C routines in Helen't example
differed from the archive example. and that neither example gave the
line number in function ncx_put_float_double() in which the exception
occurred.

One thing is to ensure that the library is compiled with debugging
turned one (i.e. "-g" option) so that a stack trace of the exception
shows the exact line number.

In file "libsrc/ncx_cray.c" (which, presumably, is being used by the
T3E), there's some conditional code that's activated by floats and
doubles being equal:

    #if SIZEOF_FLOAT == SIZEOF_DOUBLE && FLT_MANT_DIG == DBL_MANT_DIG
            return put_ix_float(xp, (float *)ip);
    #else

where "ip" is a pointer to a double. Is this code active on the T3E? If
the code is modified so that the above section is not active but the
"#else" is (which converts the double value to a float), does the
problem still occur?

Regards,
Steve Emmerson   <http://www.unidata.ucar.edu>