• Home
Name Date Size #Lines LOC

..--

bench/03-May-2024-1,3351,084

cmake/03-May-2024-4839

include/03-May-2024-795388

test/03-May-2024-4,4733,992

.gitignoreD03-May-2024180 2017

Android.bpD03-May-20242.7 KiB130118

CMakeLists.txtD03-May-20247.6 KiB182154

LICENSED03-May-20241.1 KiB127

METADATAD03-May-2024423 2019

MODULE_LICENSE_MITD03-May-20240

OWNERSD03-May-2024219 1211

README.mdD03-May-20241.3 KiB2114

TEST_MAPPINGD03-May-2024393 2625

configure.pyD03-May-20242.8 KiB6647

confu.yamlD03-May-2024191 88

README.md

1# FP16
2Header-only library for conversion to/from half-precision floating point formats
3
4## Features
5
6- Supports IEEE and ARM alternative half-precision floating-point format
7    - Property converts infinities and NaNs
8    - Properly converts denormal numbers, even on systems without denormal support
9- Header-only library, no installation or build required
10- Compatible with C99 and C++11
11- Fully covered with unit tests and microbenchmarks
12
13## Acknowledgements
14
15[![HPC Garage logo](https://github.com/Maratyszcza/PeachPy/blob/master/logo/hpcgarage.png)](http://hpcgarage.org)
16[![Georgia Tech College of Computing logo](https://github.com/Maratyszcza/PeachPy/blob/master/logo/college-of-computing.gif)](http://www.cse.gatech.edu/)
17
18The library is developed by [Marat Dukhan](http://www.maratdukhan.com) of Georgia Tech. FP16 is a research project at [Richard Vuduc](http://vuduc.org)'s HPC Garage lab in the Georgia Institute of Technology, College of Computing, School of Computational Science and Engineering.
19
20This 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.
21