Google Git
Sign in
android / platform / external / FP16 / 5a4d587c7896c926c644bccc89445bb4aac24525
commit5a4d587c7896c926c644bccc89445bb4aac24525[log] [tgz]
authorFlorian Mayer <[email protected]>Sat Mar 19 01:31:07 2022 +0000
committerAutomerger Merge Worker <[email protected]>Sat Mar 19 01:31:07 2022 +0000
tree2c085b686688abfc7135e4727b9ff04a6e23a4b6
parent385ac0c600f46730c209a2caf39ee71611150380 [diff]
parenta2d590fa0617fcd0eba86d5545c4242ee3f37c24 [diff]
Add FP16 to hwasan-postsubmit am: cad49b6d5a am: d91445dc42 am: f6ebfdb852 am: 8217ca7fcb am: a2d590fa06

Original change: https://android-review.googlesource.com/c/platform/external/FP16/+/2032272

Change-Id: I18bc24b553d222512070f12f6a616ecba33b4892
tree: 2c085b686688abfc7135e4727b9ff04a6e23a4b6
  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.