[mvapich-discuss] Configure bug in romio

[Glaesemann, Kurt R]

Worked for me.

Kurt Glaesemann, Ph.D.
Senior IT Engineer and Physical Chemist
P.S.  Brief and misspelled, because sent from iPhone.

> On Aug 24, 2017, at 6:34 PM, Panda, Dhabaleswar <panda at cse.ohio-state.edu> wrote:
> 
> Hi, Rob - Thanks for coming-up with this patch. 
> 
> Hi, Kurt - Please let us know whether this works out. After that, we will take-in the patch. 
> 
> Thanks, 
> 
> DK
> 
> 
> ________________________________________
> From: mvapich-discuss-bounces at cse.ohio-state.edu on behalf of Latham, Robert J. [robl at mcs.anl.gov]
> Sent: Thursday, August 24, 2017 10:33 AM
> To: Kurt.Glaesemann at pnnl.gov; mvapich-discuss at cse.ohio-state.edu
> Subject: Re: [mvapich-discuss] Configure bug in romio
> 
>> On Thu, 2017-08-24 at 13:59 +0000, Glaesemann, Kurt R wrote:
>> This file:
>> src/mpi/romio/confdb/aclocal_romio.m4
>> 
>> Has the following code in it (which is then used to generate the
>> configure script).  The last line needs to add $FFLAGS so that
>> compiler options the effect KIND will be in effect.  This is useful
>> for the nagfor -kind=byte option, which seems to work everywhere else
>> in the MVAPICH2 configure process.
> 
> There were two places where ROMIO's configure invoked the fortran
> comnpiler this way.  Can you try out the attached patch and confirm it
> works for you?
> 
> 
>> 
>> 
>> AC_MSG_CHECKING([for Fortran 90 KIND parameter for 8-byte integers])
>> cat <<EOF > conftest.$ac_f90ext
>>      program main
>>      integer i
>>      i = selected_int_kind(16)
>>      open(8, file="conftest.out", form="formatted")
>>      write (8,*) i
>>      close(8)
>>      stop
>>      end
>> EOF
>> if test -z "$FC" ; then
>>   FC=f90
>> fi
>> if $FC -o conftest$EXEEXT conftest.$ac_f90ext >/dev/null 2>&1 ; then
>> 
>> 
>> Kurt
>> 
>> ______________________
>> 
>> Kurt R. Glaesemann, Ph.D.
>> 
>> Senior IT Engineer and Physical Chemist
>> High Performance Computing and Cloud Services
>> 
>> _______________________________________________
>> mvapich-discuss mailing list
>> mvapich-discuss at cse.ohio-state.edu
>> http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss