Advanced Topics in HDF5

The following topics are broadly applicable in HDF5 applications.

• Dynamic Plugins in HDF5 - Describes the use of dynamic plugins, including dynamically loaded filters, in HDF5. These plugins need not be registered with the HDF5 library.
• Using the Direct Chunk Write Function - Describes another way that chunks can be written to datasets
• Copying Committed Datatypes with H5Ocopy - Describes how to copy to another file a dataset that uses a committed datatype or an object with an attribute that uses a committed datatype so that the committed datatype in the destination file can be used by multiple objects
• HDF5 Metadata - Provides a comprehensive overview of the types of metadata used in HDF5
• HDF5 Data Flow Pipeline for H5Dread - Describes data flow when reading raw data from an HDF5 dataset
• Virtual Dataset - VDS - Describes how to present data stored in several HDF5 datasets and files as a single dataset and access the data using HDF5 APIs (HDF5-1.10)

The following topics are of interest in certain limited situations such as when space utilization or performance must be carefully tuned.

• Collective Metadata I/O  - Describes optimizing HDF5 by collecting metadata I/O (HDF5-1.10)
• Metadata Caching in HDF5 - Provides detailed information for managing and tuning HDF5 metadata caching
• HDF5 File Image Operations   - Describes how to work with HDF5 files in memory. Disk I/O is not required when file images are opened, created, read from, or written to
• Modified Region Writes   - Describes how to set write operations for in-memory files so that only modified regions are written to storage. Available when the Core (Memory) VFD is used.
• Freeing Memory Allocated by the HDF5 Library   - Describes how inconsistent memory management can cause heap corruption or resource leaks and possible solutions
• File Space Management  -  Describes mechanisms to manage space in an HDF5 file (HDF5-1.10)
• Fine-tuning the Metadata Cache  - Describes how to provide better control of the metadata cache (HDF5-1.10)
• Enabling a Strict Consistency Semantics Model in Parallel HDF5   - Describes how to improve file access consistency semantics with the Parallel HDF5 library
• Page Buffering - Describes how to reduce the number of small accesses in HDF5 by buffering metadata and raw data accesses (HDF5-1.10)
• Partial Edge Chunks  - Describes how to control the filtering of partial edge chunks to improve performance when extending datasets (HDF5-1.10)
• Single Writer/Multiple-Reader - Enables writing to a file while multiple readers are accessing it (HDF5-1.10)
• Thread-safe HDF5 - Information on thread-safety and concurrent access in HDF5