I get the same error on a 32 bits architecture if I use the tuned
collectives.
mpirun -np 4 -mca btl tcp,self -mca coll tuned,basic
src/MPI_Allreduce_user_c
dump tons of MPITEST error (0): i=64, int value=4, expected 1
.
If I disable the tuned module:
> mpirun -np 4 -mca btl tcp,self -mca coll
Hi Graham,
On Jan 14, 2006, at 2:07 PM, Graham E Fagg wrote:
Hi all,
whatever this fixed/changed, I no longer get corrupted memory in the
tuned data segment hung off each communicator... ! I'm still
testing to
see if I get TimPs error.
G
On Sat, 14 Jan 2006 bosi...@osl.iu.edu wrote:
Auth
Graham,
with the trunk r8695 I no longer get these errors.
Tim
Quoting Graham E Fagg :
> Hi Tim
> I can get an error but not quite the same as yours. In my case I
> get a
> segfault as someone corrupts the memory attached to a communicator
> (data
> segment). Looks like a possible inplace err
Hi Tim
I can get an error but not quite the same as yours. In my case I get a
segfault as someone corrupts the memory attached to a communicator (data
segment). Looks like a possible inplace error. Expect a fix shortly.
G.
On Tue, 10 Jan 2006, Tim Prins wrote:
Graham,
It works properly if
Graham,
It works properly if I select the basic coll component. Anyways, here is
the output you requested. The full output is about 140MB, so I killed it
before it finished...
Tim
Quoting Graham E Fagg :
> Hi Tim
> nope, can you rerun with mpirun -np 4 -mca coll_base_verbose 1
>
> and emai
Hi Tim
nope, can you rerun with mpirun -np 4 -mca coll_base_verbose 1
and email me the output?
Thanks
G
On Tue, 10 Jan 2006, Tim Prins wrote:
Hi everyone,
I have been playing around with Open-MPI, using it as a test bed for
another project I am working on, and have found that on the intel te
Hi everyone,
I have been playing around with Open-MPI, using it as a test bed for
another project I am working on, and have found that on the intel test
suite, ompi is failing the MPI_Allreduce_user_c,
MPI_Reduce_scatter_user_c, and MPI_Reduce_user_c tests (it prints
something like MPITEST error