Skip to end of metadata
Go to start of metadata

Project Summary


The project is currently in collaboration with Diamond Light Source . DLS had code for loading a XML layout in their parallel HDF5 writer plugin and said they were planning to move the code to the Area Detector version. This was not enough to have the HDF5 writer plugin work with the current experiments/requirements, so updates to the code will be done by ANL. The goal is the same as the original.


This project is to design and build an HDF5 Writer Plug-in for Area Detector that incorporates features from the NDFileNexus, NDFileHDF5XML, and NDFileHDF5 Writer Plug-ins. The NDFileNexus Writer Plug-in saves attribute data in the Nexus format only and does not support HDF5 chunking or compression. The NDFileHDF5XML Writer Plug-in allows attributes and the file layout to be specified in a XML file, and supports a statically defined chunking scheme. The Diamond Writer Plug-in supports chunking, compression, and appending images to a file, but has a static form of saving attribute data. The goal is to create a single unified HDF5 Area Detector Writer Plug-in that is distributed as part of the main Area Detector distribution, and that features configurable chunking, compression, and dynamically configurable attributes and file layout. It should run on Linux, OS X, and Windows.


The plug-in would allow more control over saving experiment data and would give the user the flexibility of picking the attributes they want to save with the data. Certain analysis schemes require data to be read in a different order than it was written. Configurable chunking will enable better read optimization. Distributing a single plug-in with the main Area Detector distribution would save operations effort during configuration and set up.

Project Management and Issue Tracking

Jira is used for project management and issue tracking.

Source Code

Source code for the project is available in its GIT repository at


Meeting Minutes

HDF5 Chunking Benchmarks

HDF5 Attributes Config

Original Deliverables ( Deprecated since DLS collaboration )

  • No labels