On Thu, 8 May 2003, Michael Madore wrote:
> I just ran lamtests-7.0b9 against LAM 7.0b9 using Bproc. All of the
> tests passed with the following exceptions:
>
> env/procname
> env/init_thread_cxx
> env/is_thr_main_cxx
> env/query_thread_cxx
We actually released b10 last night a few hours after b9 because of
exactly these problems (it was a minor error in the C++ bindings cleanup).
> With lamtests-7.0b8 and LAM 7.0b8 only the only test that fails is
> env/procname, and it fails in the same manner as b9.
>
> [**ERROR**]: LAM/MPI MPI_COMM_WORLD rank 0, file procname.c:74:
> ERROR, processor name = 0, should be n0
> [**ERROR**]: LAM/MPI MPI_COMM_WORLD rank 1, file procname.c:74:
> ERROR, processor name = 1, should be n1
> [**ERROR**]: LAM/MPI MPI_COMM_WORLD rank 2, file procname.c:74:
> ERROR, processor name = 2, should be n2
> [**ERROR**]: LAM/MPI MPI_COMM_WORLD rank 3, file procname.c:74:
> ERROR, processor name = 3, should be n3
Ah -- this is interesting. Hadn't tested this on a bproc platform; I'll
have to have a look at this. Even so, it's a minor warning and shouldn't
cause any problems in normal MPI programs.
Thanks for the bug report!
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
