尽管使用了随机种子，随机数仍然保持不变 [英] Random numbers keep coming out the same, despite random seed being used

问题描述

我有以下一小段代码：

I have the following small piece of code:

  REAL(8)       :: x
  INTEGER       :: i

  call system_clock(i)
  WRITE(*,*) 'cpu time', i
  CALL random_seed(i)

  CALL random_number(x)
  WRITE(*,*) 'uniform RandVar', x

CPU时间为工作正常，但每次运行这个时，我都会得到相同的制服 RandVar number = 0.99755959009261719 ，几乎就像 random_number 正在使用一遍又一遍的相同默认种子，并忽略随机种子。

CPU time is working fine, but every time I run this I get the same uniform RandVar number = 0.99755959009261719, almost like random_number is using the same default seed over and over again and ignoring random seed.

我做错了什么？

What am I doing wrong?

推荐答案

很好被使用：这是依赖于处理器。原因是你对 random_seed 的调用不会设置种子。

The same seed may well be being used: that is processor-dependent. The reason for this is that your call to random_seed is not setting the seed.

使用

CALL random_seed(i)

参数 i 不是（ intent（in））种子，而是（ intent（out））处理器使用的种子大小。这个调用就像

the argument i is not the (intent(in)) seed, but is the (intent(out)) size of the seed used by the processor. This call is like

CALL random_seed(SIZE=i)  ! SIZE is the first dummy argument

要设置种子，您需要明确地与<$ c $相关联c> PUT 伪参数：调用random_seed（put = seed）。这里的种子是大小至少 n 的一级数组，其中 n - 同样取决于处理器 - 是大小由调用random_seed（size = n）给出。您的电话 i 保留此值。

To set the seed you need to explicitly associate with the PUT dummy argument: call random_seed(put=seed). Here the seed is a rank 1 array of size at least n where n - again processor-dependent - is the size given by call random_seed(size=n). From your call i holds this value.

详细信息在F2008的13.7.136中给出。

Full details are given in 13.7.136 of F2008.

生成生成器的常用方法是：

A common way to seed the generator is:

integer, allocatable :: seed(:)
integer size

call random_seed(size=size)
allocate(seed(size))
! set seed(:) somehow
call random_seed(put=seed)

设置 seed 并不是一个简单的过程。我在这里没有解决如何做到这一点，但详细信息可以在这个其他问题的答案中找到。

Setting seed appropriately is not a simple process. I don't address how to do that here, but detail can be found in answers to this other question.

在评论中提到的 srand（）的使用是非标准的。

Use of srand(), which is mentioned in the comments, is non-standard.

这篇关于尽管使用了随机种子，随机数仍然保持不变的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！