Changes to Random Number Generation between r4 and r8
'''Using a Computer to Generate Random Numbers'''
My algorithm: Uses low-order bits from high precision of timing trivial programs to frequently seed the [Psuedo-Random Number Generator]. The low-order bits are high in entropy and affected by external sources such things as CPU load, CPU temperature (CPU speed varies slightly with CPU temperature), clock variance, and other unpredictable stimulus.
(blank line)
There is no seed (or rather, it is self-seeding) needed to prime the routines, and the results are not predictable given previous results.
(blank line)
Sample C implementation [http://www.rkeene.org/projects/info/resources/rng/rng.c]
(blank line)
----
Diehard Battery of Tests [http://www.stat.fsu.edu/pub/diehard/] for determining the quality of random numbers.
(blank line)
"ENT" random number testing program [http://www.fourmilab.ch/random/] [http://www.rkeene.org/projects/info/resources/rng/ent-0.0.0.tar.gz]
Only in r4
Only in r8 --> Modified slightly between r4 and r8