Re: [ESPResSo-users] Mac OS - segmentation fault and other strange error

[Marcello Sega]

Dear Paul,

a segmentation fault means the code is accessing either a non-existing
or reserved part of the memory, is not related to the limits set on
your machine.

In principle a segfault should not happen, as every possible "mistake"
in your script should in principle be caught by espresso. This is
likely a bug which surfaced because of using a different compiler &
OS.

Send please around a minimal script that produces the segfault, and
let us know the version of espresso you are using.

Best,

Marcello




On Sat, Jan 16, 2016 at 5:21 PM, Paul Peterson
<address@hidden> wrote:
> Hello everybody,
>
> I know this is not really an Espresso question (i.e. one related to its
> scientific use) but I'm confident you may be of help.
>
> I managed to compile Espresso on a Mac (El Capitan) and:
>
> 1) I got a segmentation fault:11 which I'd never gotten on the (far less
> performing) Linux machine I'd installed Espresso on in the first place.
> Since I've read that indicates a stack overflow, I checked my stack size
> with ulimit. Result: unlimited.
>
> 2) Then, I just reduced the size of the problem - it used to handle
> 40-monomer polymer, I set the length to 20. That did the trick... with the
> stack. The simulation went past the previous blocking point, and it was
> running visibly faster than under my poor Linux laptop. But immediately
> afterwards, I got a 'bond broken' error I didn't get when running the same
> (exactly the same) script on the other machine.
>
> Any suggestion would be highly appreciated.
>
> Thanks in advance,
>
> Paul



-- 
University of Vienna, Institute of Computational Physics