Christian,

Try switching off multi-file optimization, which is one of the
"complaints" you get from the compiler, if you have -ip or -ipo (or
whatever is equivalent) set to on. My compilation was successful using
more or less the same versions of intel products you are using. If
this doesn't work, I can only think of a problem using mpich 1 - I
have used the latest version of mpich 2 and everything went fine.

Marcos

On Mon, Oct 26, 2009 at 7:45 PM, Christian helman
<[email protected]> wrote:
> Hola Siesteros!
>
> Previously I sent a mail with subject "same structure - different
> siesta version - different forces". By then I had been compiled the
> new version of siesta (3.0-b) but the results were quite different
> from the previous versions (this is why the subject). Then Marcos and
> Eduardo gave me a lot of tips (Thanks again), one of them was check
> that everything is compile with the SAME fortran version! So, now I
> have my MKL's (version 10.0.1.014) compiled with intel fortran (ifort)
> 10.1.13 (all 64bits) and mpich 1.2.7 , then, when I try to compile
> Siesta I get the next error:
>
>>              fortcom: Severe: **Internal compiler error: internal abort** 
>> Please report this error along with the circumstances
>>              in which it occurred in a Software Problem    Report.  Note: 
>> File and line given may not be explicit cause of this error.
>
>>              ifort: error #10014: problem during multi-file optimization 
>> compilation (code 3)
>>              make: *** [siesta] Error 3
>
> Of course I change in arch.make the optimization option from -O3 to
> -O2 to -O0, and nothing change.
> Anybody compile siesta3.0 with this version of mkl and fortran??
> any thought? any tip?
> Thanks in advance
>
>
> Lic. Christian Santiago Helman
> Gerencia de Investigaciones y Aplicaciones
> Centro Atómico Constituyentes-CNEA
> Argentina
> Tel:++54-11-6772-7102
>

Responder a