On Monday, October 12, 2015 at 9:52:28 AM UTC-4, Takeshi wrote: > > (please note that a pair of Sobol numbers are used for each iteration, so > i/10^7 = 2140 roughly > corresponds to 2^32-1.) Although the webpage of Sobol.jl says that > Mersenne Twister will be used after the 2^32-1 samples, > do I need to switch manually to the built-in random number generator in > Julia? Or am I doing something wrong > about the treatment of Integer? I am using Linux 64bit and assuming that > integer literals is of Int64. >
It looks like this is a bug in the most recent version; you shouldn't need to switch manually. Can you please file an issue?
