Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / FP16 / 114872d6099f8452185453a1d5f42f912a3f4177
commit114872d6099f8452185453a1d5f42f912a3f4177[log] [tgz]
authorIan Hua <ianhua@google.com>Wed Feb 02 14:14:35 2022 +0000
committerPresubmit Automerger Backend <android-build-presubmit-automerger-backend@system.gserviceaccount.com>Wed Feb 02 14:14:35 2022 +0000
tree902ebbb006074e18653102fdd5be27b04e0d0a88
parent5484dbfabb408ee81ef8615182c9cc43ed9d885a [diff]
parent72cdf5e49b87f3f323fd8bdf97d6c524d6275fab [diff]
[automerged blank] Add host_supported, min_sdk_version and apex_available configs to Android.bp for FP16. 2p: 72cdf5e49b

Blank merge reason: Change-Id Ia8ac144df6a18f17cc9c181ed8618a5d365abaa6 with SHA-1 80907cb9be is in history

Original change: https://googleplex-android-review.googlesource.com/c/platform/external/FP16/+/16771984

Bug: 196532440
Change-Id: Ide4d3f46f2e98ae6ae37a1fd1c2b68a1832f69d9
tree: 902ebbb006074e18653102fdd5be27b04e0d0a88
  1. bench/
  2. cmake/
  3. include/
  4. test/
  5. .gitignore
  6. Android.bp
  7. CMakeLists.txt
  8. configure.py
  9. confu.yaml
  10. LICENSE
  11. METADATA
  12. MODULE_LICENSE_MIT
  13. OWNERS
  14. README.md
  15. TEST_MAPPING
README.md

FP16

Header-only library for conversion to/from half-precision floating point formats

Features

  • Supports IEEE and ARM alternative half-precision floating-point format
    • Property converts infinities and NaNs
    • Properly converts denormal numbers, even on systems without denormal support
  • Header-only library, no installation or build required
  • Compatible with C99 and C++11
  • Fully covered with unit tests and microbenchmarks

Acknowledgements

HPC Garage logo Georgia Tech College of Computing logo

The library is developed by Marat Dukhan of Georgia Tech. FP16 is a research project at Richard Vuduc's HPC Garage lab in the Georgia Institute of Technology, College of Computing, School of Computational Science and Engineering.

This material is based upon work supported by the U.S. National Science Foundation (NSF) Award Number 1339745. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of NSF.