def double_float_random_series (   self,
  count 
)

Definition at line 160 of file r250.py.

00160                                                :
00161         "return a series of 52-bit significance double-precision randoms on [0,1)."
00162         q1 = self.fast_random_series(2*count) #take bits from the random pool
00163         q1[1::2] &= self.mask32#mask off low bits to prevent rounding when combined to 52-bit mantissas
00164         q1=q1*self.single_floatscale #convert results to doubles scaled on [0,1)        
00165         #now, for any generator with two-point correlations, or unknown bit lengths, this would fail horribly, but this class is safe.
00166         q1[1::2]*=self.single_floatscale #rescale LSBs 
00167         q1[0::2]+=q1[1::2] #and combine with MSBs
00168         return q1[::2]
00169 
class r250(ran_shift):


Generated on Wed Nov 21 10:18:33 2007 for analysis by  doxygen 1.5.4