The ABINIT Users Mailing List ( CLOSED )

[Steven Homolya <steven.homolya@sci.monash.edu.au>]

I don't have a solution, sorry; just replying to confirm that I am having the
same or very similar problems with pathscale + opteron + MPI combination. I'm
going to try with g95 next.

Steve

fer@tiziano.phys.washington.edu wrote:
>   I am trying to compile abinip on an Opteron cluster (the jacquard machine
> at NERSC, in case anyone knows it) with the PathScale f90 compiler and the
> mvapich version of MPI.
>
>   I get the code to compile, but when I try to calculate the response
> function I get the following error:
>
> **** Segmentation fault!  Fault address: (nil)
[...]

> **** Segmentation fault!  Fault address: 0xc5e8
[...]

> Fault address is 4143640 bytes below the first valid
> mapping boundary, which is at 0x400000.
[...]

> Anybody has any experience with this combination of compiler and mpi?

--
Steven Homolya
Department of Physics
La Trobe University, VIC 3086
Australia
Tel: +61 3 9479 1430