segmentation fault on a water example

Axel akoh... at gmail.com
Tue Mar 10 15:00:19 UTC 2009


> Any ideas on what should be changed in order to get cp2k compile with
> "stronger" optimizations and without changing the compiler? (maybe I'm
> asking for to much).
> If not I'm going to try what standing on one foot as Teodoro suggested :).

but that has more chances of helping you. trying to identify which
pieces
of a code need to be compiled with or without full optimization to
make it
work with a broken compiler is a futile undertaking, you never know
why
something is breaking and people here will get very annoyed by reports
of "bugs" that are actually compiler generated. please note, that you
got
lucky to find a case where you get a segfault. what about other cases,
where
you just get an incorrect result, sometimes only intermittently.

in my experience the intel compilers on itanium are "more broken"
than on other platforms. i attribute this to the fact, that with
an EPIC architecture, the compiler has to do a lot of the work that
is done on x86-like platforms within the CPU.


cheers,
    axel.
>
> Thanks a lot
> Marius
>



More information about the CP2K-user mailing list