From a75effa9489e41915b624084c61fba144003b448 Mon Sep 17 00:00:00 2001
From: Robert David Grant <rgrant@enthought.com>
Date: Tue, 24 Jun 2014 18:10:45 -0500
Subject: [PATCH] Add a simple test of one's h5py installation.

---
 utils/test_h5py.py | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)
 create mode 100644 utils/test_h5py.py

diff --git a/utils/test_h5py.py b/utils/test_h5py.py
new file mode 100644
index 00000000..f2949560
--- /dev/null
+++ b/utils/test_h5py.py
@@ -0,0 +1,37 @@
+"""
+Simple test of a parallel build of h5py (from h5py's documentation).
+
+http://h5py.readthedocs.org/en/latest/mpi.html#using-parallel-hdf5-from-h5py
+
+If you've built h5py properly against a parallel build of hdf5, you should be
+able to run this code with::
+
+    $ mpiexec -n 4 python test_h5py.py
+
+and then check the output with `h5dump`::
+
+    $ h5dump parallel_test.hdf5
+    HDF5 "parallel_test.hdf5" {
+    GROUP "/" {
+    DATASET "test" {
+        DATATYPE  H5T_STD_I32LE
+        DATASPACE  SIMPLE { ( 4 ) / ( 4 ) }
+        DATA {
+        (0): 0, 1, 2, 3
+        }
+    }
+    }
+    }
+"""
+
+from mpi4py import MPI
+import h5py
+
+rank = MPI.COMM_WORLD.rank  # The process ID (integer 0-3 for 4-process run)
+
+f = h5py.File('parallel_test.hdf5', 'w', driver='mpio', comm=MPI.COMM_WORLD)
+
+dset = f.create_dataset('test', (4,), dtype='i')
+dset[rank] = rank
+
+f.close()