Fallback to the regular pages if huge page allocation is set to auto
and it was not possible to allocate requested amount of memory with
the hugepages.
Signed-off-by: Alex Mikheev <alexm@mellanox.com>
Change the default enable configure option XRC to disabled. If a user want's
to give it a try they have to explicitly ask for it.
Modify the configury help message to indicate it is not enabled by default.
Related to #3890Fixes#3969
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Every modern compiler supports either inline assembly or builtin atomic
operations. Because of this it is time to delete all the code associated
with pre-built atomics.
This commit also clean out the DEC and XLC asm checks. Neither check
does anything and the XLC compiler supports GCC ASM.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
always re-initialize vmhole *before* mca_base_component_var_register()
otherwise the vmhole gets NULL'ified if orte is initialized a second time.
that typically occurs when Open MPI is configure'd with --disable-dlopen
and the app does MPI_T_init_thread(); MPI_T_finalize(); MPI_T_init_thread();
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
It turns out that the approach of having the HNP do the
fork/exec of MPI ranks on the head node in a SLURM environment
introduces problems when users/sysadmins want to use the SLURM
scancl tool or sbatch --signal option to signal a job.
This commit disables use of the HNP fork/exec procedure when
a job is launched into a SLURM controlled allocation.
update NEWS with a blurb about new ras framework mca parameter.
related to #3998
Signed-off-by: Howard Pritchard <hppritcha@gmail.com>
This removes a copy-and-paste error where we were setting the
OPAL_ASM_SYNC_HAVE_64BIT more than once.
References #3993. Close when on master and v3.0.x.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
There was a typo in the shmem_g c11 generic interface
in shmem.h.in
Thanks to @nspark for reporting the problem and
specifying the fix.
Fixes#3968
Signed-off-by: Howard Pritchard <hppritcha@gmail.com>
Ignore errors caused by remote side having exited when closing CUDA IPC mappings.
openmpi/ompi#3244
Signed-off-by: Sylvain Jeaugey <sjeaugey@nvidia.com>
The `ompi_comm_set` function never sets `NULL` to its first argument
`ncomm`. So `NULL` check is unnecessary in its callers. Furthermore,
`NULL` check may obscure a real return code when an error occurs
if the variable is initialized to a `NULL` value.
Also, `NULL` check is added in the `ompi_comm_set` function to
avoid segmentation fault in an out-of-memory condition.
Signed-off-by: KAWASHIMA Takahiro <t-kawashima@jp.fujitsu.com>
Per discussion at the Summer 2017 developers meeting, generate
the AUTHORS list at make dist time, rather than trying to
keep it up to date and merge on the branches by hand. While
most of the data is generated from git, the organization list
was maintained by hand. The general feeling at the meeting was
that the organization list was not adding value and there were
concrete cases where it involved much chasing by the RMs, so
it has been removed.
Signed-off-by: Brian Barrett <bbarrett@amazon.com>
This changes the default to 0, to avoid yields during progress in srun.
In mpirun, ompi_mpi_yield_when_idle is set to 1 if oversubscribed
otherwise 0. But the default is 1 though, and it is used in srun.
Now srun and mpirun have the same latency in non-oversubscribed cases.
Signed-off-by: Piotr Lesnicki <piotr.lesnicki@atos.net>
