[mpich-discuss] Help with mpich2 1.3a1 on i7-920

[Rajeev Thakur]

This typically happens when you pass the same buffer as input and output
argument to an MPI collective function. You need to either use separate
buffers or use the MPI_IN_PLACE option supported by the collectives.

Rajeev
 

> -----Original Message-----
> From: mpich-discuss-bounces at mcs.anl.gov 
> [mailto:mpich-discuss-bounces at mcs.anl.gov] On Behalf Of 
> Gaetano Bellanca
> Sent: Wednesday, February 17, 2010 6:26 PM
> To: mpich-discuss at mcs.anl.gov
> Subject: [mpich-discuss] Help with mpich2 1.3a1 on i7-920
> 
> Hi,
> 
> I have compiled the new 1.3a1 version of mpich2 on a (single) Intel 
> i7-920 (ubuntu 9.10 - Intel Fortran 11.1).
> This is a new machine just installed today, and I'm running the first 
> tests.
> I have successfully compiled my simulator but, launching the 
> simulation 
> with:
> mpiexec -n 2 -f hosts -bootstrap ssh my_simulator
> 
> I have the message:
> Assertion failed in file helper_fns.c at line 337: 0
> memcpy argument memory ranges overlap, dst_=0x2636cf0 
> src_=0x2636cf0 len_=4
> 
> the host file contains two times the name of the computer.
> 
> The same code runs on a different computer with mpich2 1.2.1.
> Is it something related to the 1.3a1 version, to the use of 
> hydra, to he 
> new architecture I'm using?
> 
> Any idea?
> 
> Regards.
> 
> Gaetano
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>