commit | 5e2bd58cc5bfa0738fae08e388eec0ecfba041c4 | [log] [tgz] |
---|---|---|
author | Marat Dukhan <marat@fb.com> | Sun Sep 23 19:51:07 2018 |
committer | Marat Dukhan <marat@fb.com> | Sun Sep 23 19:54:33 2018 |
tree | 4a14aeabe5abef32bec1de0a47bcd88769d4b606 | |
parent | 4b37bd31c9cc1380ef9f205f7dd031efe0e847ab [diff] |
Fix compilation in strict C++11/C++14 mode Hexadecimal floating-point literals is a C99 feature. gcc/icc/clang support hex-float literals in C++ as an extension, and fail to compile such code in strict C++11 or C++14 mode (i.e. -std=c++11) This patch adds a work-around for when modes where hex-float literals are not supported..
Header-only library for conversion to/from half-precision floating point formats
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.