]> https://gitweb.dealii.org/ - dealii.git/commitdiff
Improve the documentation
authorDaniel Garcia-Sanchez <daniel.garcia-sanchez@insp.upmc.fr>
Tue, 9 Oct 2018 14:37:55 +0000 (16:37 +0200)
committerDaniel Garcia-Sanchez <daniel.garcia-sanchez@insp.upmc.fr>
Wed, 9 Jan 2019 18:25:03 +0000 (19:25 +0100)
Thread safety documentation

Modify the hyperslab documentation

include/deal.II/base/hdf5.h

index 6b4c4bb7a564a86d90ed544607cd2764906b6465..8787327df123492eed4101169f8cd3653583b199 100644 (file)
@@ -127,7 +127,8 @@ DEAL_II_NAMESPACE_OPEN
  * ## Write a hyperslab in parallel
  * Hyperslabs are portions of datasets. A hyperslab can be a contiguous
  * collection of points in a dataset, or it can be a regular pattern of points
- * or blocks in a datataset.
+ * or blocks in a datataset. Hyperslabs are equivalent to python numpy and h5py
+ * [slices](http://docs.h5py.org/en/latest/high/dataset.html#reading-writing-data).
  *
  * See the <a
  * href="https://support.hdfgroup.org/HDF5/doc/UG/HDF5_Users_Guide-Responsive%20HTML5/HDF5_Users_Guide/Dataspaces/HDF5_Dataspaces_and_Partial_I_O.htm?rhtocid=7.2#TOC_7_4_Dataspaces_and_Databc-6">Dataspaces
@@ -300,6 +301,14 @@ DEAL_II_NAMESPACE_OPEN
  * 4 5
  * @endcode
  *
+ * # HDF5 and thread safety
+ * By default HDF5 is not thread-safe. The HDF5 library can be configured to be
+ * thread-safe, see [the HDF5
+ * documentation](https://support.hdfgroup.org/HDF5/faq/threadsafe.html). The
+ * thread-safe HDF5 version serializes the API but does not provide any level of
+ * concurrency. To achieve high parallel performance with HDF5, we advice to use
+ * HDF5 with MPI.
+ *
  * @author Daniel Garcia-Sanchez, 2018
  */
 namespace HDF5

In the beginning the Universe was created. This has made a lot of people very angry and has been widely regarded as a bad move.

Douglas Adams


Typeset in Trocchi and Trocchi Bold Sans Serif.