Random generator "periodicy" [message #7243] |
Wed, 27 August 2008 10:25 |
Ralf Kliemt
Messages: 507 Registered: May 2007 Location: GSI, Darmstadt
|
first-grade participant |
From: 141.30.85*
|
|
Hello everyone,
Lately I have seen some structures in a simulation with 10.000 events. It made me wonder how many events we can safely simulate with one seed without hitting the random number generators "periodicy" limits where we cannot count on the (pseudo-) randomness of the numbers.
One type of tool to check this might be these performance checkers which give you the amount of calls to all of your functions. Unfortunately I don't recall any of these tools names since I never used one.
Rene said e.g. that in bbrroot people use 1000 events per random seed.
Greetings from Dresden, Ralf.
|
|
|