[gmx-users] g_energy floating exception

[Erik Lindahl]

Sergio Oyama Junior wrote:

>Hi Erik,
>
>I apologize for the delay of my reply. I have been doing some tests and
>noticed some interesting aspects of this problem. Unfortunately, the
>compilation with the "-g" flags didn't result in a functional binary in my
>system. However, I have observed that the occurrence of this floating
>exception depends on the energy file size. The amazing point is that
>g_energy from v3.1.3 does work with very short or very large files. For
>example, with a file generated by 5 steps of minimization (just as a test)
>the program works. Also, it does work well with energy files resulting from
>simulations of 1 ns or higher. The exception occurs with calculations are
>done in the time range of 100 to 800 ps, as far as I have tested. Isn't it
>strange?
>
Sounds like it could be some kind of memory allocation problem, but I 
can't understand why it works on every other platform.
Still, the file you sent me - it crashed every time on 3.1.3 for you, 
right?

We could try to do it the other way around - send me your executable and 
I'll see if I can reproduce the crash
on a different version of Tru64 (too check if it's the libraries or not).

>
>
>There's another question I would like to address you. I'm having another
>kind of problem with g_disre. This program results in core dumped regardless
>the  size of the trajectory. Do you know what could be the problem?
>
Sorry, g_disre is one of the programs that I've never used... most of my 
simulations are on membranes :-)

Cheers,

Erik