Support for half-precision (16-bit) floating-point datasets #587
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
alkino
reviewed
Jun 7, 2022
alkino
reviewed
Jun 7, 2022
|
clang-format is version 12 here: https://github.com/BlueBrain/HighFive#code-formatting |
#588 begs to differ. |
Codecov Report
@@ Coverage Diff @@
## master #587 +/- ##
=========================================
Coverage ? 80.57%
=========================================
Files ? 66
Lines ? 3572
Branches ? 0
=========================================
Hits ? 2878
Misses ? 694
Partials ? 0 Continue to review full report at Codecov.
|
alkino
approved these changes
Jun 8, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Although 16-bit float is not a native C++ type, it is widely used to save runtime and disk space in situations where high precision is not needed, most notably in deep-learning applications. Python's
h5pytrivially supports the creation offloat16datasets (type "<f2"). This PR adds optional support for half-precision in HighFive.Code change summary
include/highfive/bits/H5DataType_misc.hpp, where theAtomicTypeconstructor was specialized forfloat16_tnumerical_test_typeswas augmented by addingfloat16_t(if enabled)create_dataset_half_float.cppinsrc/examplesHow to test this?
Download the (single-header)
Halflibrary from http://half.sourceforge.net/, and copyhalf.hppinto/usr/local/include/.Then build and run unit tests:
cmake .. -DHIGHFIVE_USE_HALF_FLOAT=ON make -j8 make testYou should see tests with
numerical_test_types - 13, which corresponds tofloat16_tTest System