I'm performing a lengthy computation that uses a lot of calls to rand, and I want to be able to pause the computation and resume it again later (using Storable or some other serialization). The process starts with a random seed that I provide, so that I can reproduce interesting results later. I want the results to be the same for a given initial random seed regardless of whether I paused/resumed the computation, so I need to somehow save a random seed in my serialization.

The only reasonable* solution I could think of was something along these lines:

## either load the saved random seed and state data from the
## serialization (resume a paused compuatation), or start with 
## new state data and the given random seed (start new 
## compuation)
my ($foo, $bar, $random_seed) = deserialize('saved')
                             || initial_values();

while (1) {
    srand($random_seed);

    ## some crunching on $foo, $bar, using lots of rand()

    ## decide what the next iteration's seed will be
    $random_seed = int rand( 2 << 31 );
    
    if (check_for_user_interrupt()) {
        ## save state, along with the seed for next iteration
        serialize('saved', $foo, $bar, $random_seed);
        exit;
    }
}
[download]

I use rand to generate a new seed for the next iteration, so it can be saved after any iteration. As soon as I wrote srand inside my while loop, red flags popped up in my head. The srand perldoc says:

Do not call srand() multiple times in your program unless you know exactly what you're doing and why you're doing it.

... but I wonder if I really know exactly what I'm doing ;). I found at least one post regarding multiple srand calls, but in that case the poster was using something like time ^ $$ to re-seed within a very fast loop, and expecting different random output for each loop.

What I would really like to know is in what situations is ok to call srand multiple times? Is using a call to rand for a seed considered bad? I can't use any external entropy to re-seed, because I want all the results to be based on the original random seed supplied at the beginning of the computation. Any thoughts on the serialization of randomly-generated state data like this?

*: In the worst case I could always wrap rand, save the original random seed, and simply save the number of times rand was called. Then to resume again, I just reseed, and call rand in void context a whole slug of times. I'd rather not do this!

blokhead

---

---

• Are you posting in the right place? Check out Where do I post X? to know for sure.
• Posts may use any of the Perl Monks Approved HTML tags. Currently these include the following:

  <code> <a> <b> <big> <blockquote> <br /> <dd> <dl> <dt> <em> <font> <h1> <h2> <h3> <h4> <h5> <h6> <hr /> <i> <li> <nbsp> <ol> <p> <small> <strike> <strong> <sub> <sup> <table> <td> <th> <tr> <tt> <u> <ul>

• Snippets of code should be wrapped in <code> tags not <pre> tags. In fact, <pre> tags should generally be avoided. If they must be used, extreme care should be taken to ensure that their contents do not have long lines (<70 chars), in order to prevent horizontal scrolling (and possible janitor intervention).
• Want more info? How to link or How to display code and escape characters are good places to start.