# half_precision

**Repository Path**: qiyixuan/half_precision

## Basic Information

- **Project Name**: half_precision
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-10-18
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Half Precision Type (float16)

A library for 16-bit half-precision float point type. All the calculation will
be converted to 32-bit float and then downgrade to 16-bit one.

## Requirement
- CMake 3.1+
- A C++ compiler that supports C++11 standard

## Installation & Test:
```Shell
mkdir build && cd build
cmake ..
make
make install
# test if needed
make test
```

## Usage

### Source file
```C++
...
// main.cpp
// Include the header
#include <float16.hpp>
...
// Define a value
float16 val = 1.0;
// Do what you want
...
```

### Compilation
```Shell
${CXX} -std=c++11 main.cpp -I/PATH/TO/float16.hpp -L/PATH/TO/float16.lib -lfloat16
```

## TODO
- Performance optimization
- Standard compatibility