Bug: 148976166

Clone this repo:
  1. a31640c Upgrade FP16 to 4dfe081cf6bcd15db339cf2680b9281b8451eeb3 am: 7cea9a0f20 am: 6321f9ece6 am: 169fc31aa8 am: 654ee0f4d7 by Haibo Huang · 9 months ago master
  2. 654ee0f Upgrade FP16 to 4dfe081cf6bcd15db339cf2680b9281b8451eeb3 am: 7cea9a0f20 am: 6321f9ece6 am: 169fc31aa8 by Haibo Huang · 9 months ago
  3. 169fc31 Upgrade FP16 to 4dfe081cf6bcd15db339cf2680b9281b8451eeb3 am: 7cea9a0f20 am: 6321f9ece6 by Haibo Huang · 9 months ago
  4. 6321f9e Upgrade FP16 to 4dfe081cf6bcd15db339cf2680b9281b8451eeb3 am: 7cea9a0f20 by Haibo Huang · 9 months ago android-r-beta-2 android-r-beta-3
  5. 7cea9a0 Upgrade FP16 to 4dfe081cf6bcd15db339cf2680b9281b8451eeb3 by Haibo Huang · 9 months ago

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.