Merge branch 'mpi' into singleenv

Author: bcumming

docs/cluster-config.md

@@ -46,7 +46,7 @@ packages:
     openmpi:   ...
 ```
 
-??? example "example `network.yaml` for grace hopper"
+??? example "example `network.yaml` for Grace Hopper"
     * The `specs` field for `mpi:cray-mpich:specs` and `mpi:openmpi:specs` fields set different default `libfabric` for the respective MPI distributions.
     * By default `packages:cray-mpich` and `packages:openmpi` add the `+cuda` variant as a preference to build with cuda support by default on the Grace-Hopper nodes.
         * This can be overriden by adding `~cuda` to the spec in `network:mpi` in your recipe.

docs/porting.md

@@ -201,7 +201,7 @@ The `mpi:` field is gone, and in its place there is a `network:` field.
           mpi: cray-mpich@8.1.30 +cuda
         ```
 
-    If building on a system with NVIDIA GPUs, the `+cuda` option is probably selected by default.
+    If building on a system with NVIDIA GPUs, the `+cuda` option is selected by default.
     See the `mpi:cray-mpich` and `packages:cray-mpich` fields for the `network.yaml` file in the cluster configuration to see the defaults.
 
 

docs/recipes.md

@@ -187,7 +187,7 @@ The `network` field has a field for defining MPI and additional custom package d
     ```yaml title="environments.yaml"
     network:
       mpi: openmpi@5.0.8
-      specs: ['libfabric@1.2.2']
+      specs: ['libfabric@2.2.0']
     ```
 
 It is only possible to have a single MPI implementation in an environment, specified through the `mpi` field.
@@ -231,7 +231,7 @@ See the [`network.yaml` documentation][ref-cluster-config-network] for more info
 
 !!! alps
 
-    The recommended MPI distribution on Alps is `cray-mpich`, as it is the most widely tested MPI for the libfabric/slingshot network.
+    The recommended MPI distribution on Alps is `cray-mpich`, as it is the most widely tested MPI for the libfabric/Slingshot network.
 
     OpenMPI's support for the Slingshot network is improving, however it may not be optimal for many applications, or requires more effort to fine tune.
     As such, it is recommended as an option for applications that have performance issues or bugs with cray-mpich.

stackinator/recipe.py

@@ -111,10 +111,10 @@ def __init__(self, args):
         # - meta data about mpi
         # - package information for network libraries (libfabric, openmpi, cray-mpich, ... etc)
         network_path = self.system_config_path / "network.yaml"
-        self._logger.debug(f"opening {network_path}")
         network_packages = {}
         mpi_templates = {}
         if network_path.is_file():
+            self._logger.debug(f"opening {network_path}")
             with network_path.open() as fid:
                 raw = yaml.load(fid, Loader=yaml.Loader)
                 if "packages" in raw: