Need help with FP16?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

Maratyszcza
153 Stars 40 Forks MIT License 45 Commits 8 Opened issues

Description

Conversion to/from half-precision floating point formats

Services available

!
?

Need anything else?

Contributors list

# 8,776
ml
matrix-...
fast-fo...
assembl...
27 commits
# 226,386
c-plus-...
TeX
Linux
JavaFX
1 commit
# 417
C++
faceboo...
Shell
caffe2
1 commit

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.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.