The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

Before getting an answer from someone who actualy made the compilation
on such system you can try to change the compilation options.
What happen if you configure with the following command (just for tests 
purpose)
../configure FC=ifort FCFLAGS="-O0"  ?

regards

PMA

On 4/6/07, dekoning@ifi.unicamp.br <dekoning@ifi.unicamp.br> wrote:
> Dear all,
>
> I'm still having trouble compiling Abinit 5.3.4 on an SGI Altix Itanium system using ifort 9.0.  I 
> executed "configure" without a <hostname>.ac (to use the standard options) followed by 
> "make". It produces executables but they are no good. If I execute make tests_in afterwards 
> it gives me the following:
>
> ****************************************************************************
> make basic_tests
> make[1]: Entering directory 
> `/home/esmeralda/dekoning/dekoning/abinit/abinit-5.3.4'
> cd ./tests && make tests_in
> make[2]: Entering directory 
> `/home/esmeralda/dekoning/dekoning/abinit/abinit-5.3.4/tests'
> Running built-in test 1
> *** glibc detected *** free(): invalid next size (normal): 0x60000000004875b0 
> ***
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image              PC                Routine            Line        Source
> Unknown            2000000000465C51  Unknown               Unknown  Unknown
> Unknown            20000000004666F0  Unknown               Unknown  Unknown
> abinis             4000000001F8ECE0  Unknown               Unknown  Unknown
> abinis             4000000001A351F0  Unknown               Unknown  Unknown
> abinis             4000000001A1BF90  Unknown               Unknown  Unknown
> abinis             40000000012D3C10  Unknown               Unknown  Unknown
> abinis             40000000005FA500  Unknown               Unknown  Unknown
>
> Stack trace terminated abnormally.
>
>  Status file, with repetition rate  99, status number    10
>
>   Level abinit         : call driver
>   Level driver         : enter
>  
> ********************************************************************************
>
> Evidently there is something wrong, and I suspect it may have to do with 
> compilation flags. I don't think it is a bug in the compiler. If anyone 
> managed to successfully compile Abinit on the Altix, could you please let me 
> in on the secret?
>
> Thanks!


--
Pierre-Matthieu Anglade