2

In a multithreaded program where each thread has its own state of the rand48 PRNG, what is a good relation between the seeds.

Obviously, I cannot use the same seed since then they generate the same sequences. How uncorrelated will they be for seeds {0,1,2,3} or {1,2,3,4} or whatever. I do want a fixed seed so I can repeat simulations.

user877329
  • 6,717
  • 8
  • 46
  • 88
  • I'm curious to know as well... you could use a prng whichi is somehow based on a counter, eg the one posted by djb at https://groups.google.com/forum/?hl=en&fromgroups=#!topic/sci.physics/SWhCr8SEVWw (just an example, chosen because it's fast and has longer period than rand48) – loreb Sep 12 '13 at 14:03

0 Answers0