https://github.com/pandoraeartha/highperformancerandomengine
Head only high performance pseudorandom engine base on PCG-XSH-RR. Multithreading allowed, easy to use, statistically good. 仅有头文件的随机引擎, 基于PCG-XSH-RR算法, 允许多线程, 简单易用, 高随机性.
https://github.com/pandoraeartha/highperformancerandomengine
c cpp random random-generation
Last synced: 2 months ago
JSON representation
Head only high performance pseudorandom engine base on PCG-XSH-RR. Multithreading allowed, easy to use, statistically good. 仅有头文件的随机引擎, 基于PCG-XSH-RR算法, 允许多线程, 简单易用, 高随机性.
- Host: GitHub
- URL: https://github.com/pandoraeartha/highperformancerandomengine
- Owner: PandoraEartha
- Created: 2025-03-04T10:56:22.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-04T11:50:42.000Z (over 1 year ago)
- Last Synced: 2025-03-04T12:29:05.498Z (over 1 year ago)
- Topics: c, cpp, random, random-generation
- Language: C
- Homepage:
- Size: 23.4 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
[中文介绍](https://zhuanlan.zhihu.com/p/27708881274)
# High Performance C Random Engine
Head only high performance pseudorandom engine base on PCG-XSH-RR. Multithreading allowed, easy to use, statistically good.
仅有头文件的随机引擎, 基于PCG-XSH-RR算法, 允许多线程, 简单易用, 随机性强.
## How to use
`#include "PCG32.h"`
```
PCG32Struct PCGStatus;
PCG32SetSeed(&PCGStatus,time(NULL));
double random=PCG32UniformReal(&PCGStatus,-1,999);
```
note:
1.Remember to set the seed. 记得设置种子.
2.Use its own `PCG32Struct` in each thread function and set different seed. 每个线程函数采用各自的`PCG32Struct`并且设置单独的种子
# Functions
`void PCG32SetSeed(PCG32Struct* status,long long unsigned int seed);`
Set the seed of PCG random engine and initlize
初始化随机引擎, 设置随机种子
`unsigned PCG32(PCG32Struct* status);`
Generate a unsigned type random number in range of [0,0xFFFFFFFF(4294967295)]
产生[0,0xFFFFFFFF(4294967295)]范围的`unsigned`类型的随机整数
`unsigned PCG32Uniform(PCG32Struct* status,unsigned min,unsigned max);`
Generate a unsigned type random number that obey uniform distrubution in range of [min,max]
产生[min,max]范围的`unsigned`类型的随机整数
`double PCG32UniformReal(PCG32Struct* status,double min,double max);`
Generate a double type random number that obey uniform distrubution in range of [min,max]
产生[min,max]范围的`double`类型的随机整数
`double PCG32StandardNormal(PCG32Struct* status);`
Generate a double type random number that obey standard normal distrubution
产生`double`类型的服从标准正态分布的随机数
`void PCG32UniformShuffle(PCG32Struct* status,unsigned* array,unsigned length);`
Shuffling, swaps elements in the input sequence to ensure that every element etc. can appear anywhere in the array.
洗牌, 对输入序列进行元素交换, 保证每个元素等可能的出现在数组中任意位置.
`unsigned PCG32Uniform_Strict(PCG32Struct* status,const unsigned min,const unsigned max);`
Generate a unsigned type random number that obey uniform distrubution in range of [min,max], `max` can not smaller than `min` and gap(`max`-`min`) can not be power of 2
产生[min,max]范围的`unsigned`类型的随机整数, `max`不能小于`min`且gap(`max`-`min`)不能是2的幂
# Performance
Performance test base on 13490F WSL2 Ubuntu, built using g++ 14.02. Compare with `std::default_random_engine` in C++ `random`. The result below shows the time consuming on generating 4074074037 random number on spicific distribution.
性能测试基于13490F WSL2 Ubuntu, 采用g++ 14.02编译. 与C++`random`库中的`std::default_random_engine`进行比较. 下面的结果显示了在特定分布上生成4074074037随机数所花费的时间.
## Uniform Real Distribution
`PCG32UniformReal` is 8.6399 times faster than `std::uniform_real_distribution`
`PCG32UniformReal` 比 `std::uniform_real_distribution` 快8.6399倍
## Normal Distribution
`PCG32StandardNormal` is 5.8596 times faster than `std::normal_distribution`
`PCG32StandardNormal` 比 `std::normal_distribution` 快5.8596倍
## Base Generator
`PCG32` is 1.9767 times faster than `std::default_random_engine`
`PCG32` 比`std::default_random_engine`快1.9898倍
## Uniform Unsigned
`PCG32Uniform` is 1.7713 times faster than `std::uniform_int_distribution`
`PCG32Uniform` 比 `std::uniform_int_distribution` 快1.7713倍