Information on calculator Random Number Generators from PPC Journal articles

[Namir]

(11-13-2016 10:46 PM)Dieter Wrote:  

(11-13-2016 10:09 PM)Gene Wrote:  Ok, so for newseed = FRC (PI + Seed) ^ 5 how does this look to save as much of the stack as we can?

Do we have to?

Namir, could you do a few more runs (like 100 times a sequence of 10.000 random numbers)? According to my results the ^5 version is not better than the ^3 version, as far as Chi² and autocorrelation are concerned. I may be wrong, but just to be sure...

(11-13-2016 10:09 PM)Gene Wrote:  Assumes seed is already in 00. FEEL free to condense.

Simple:

Code:

LBL R
PI
ST+ 00
X<> 00
X^2
ST* L
ST* L
X<> L
FRC
STO 00
END

But now let's wait for an MCode version of the 9821 generator using R00 so that no buffer is required (cf. the other RNG thread). I think this will be the best solution as, unlike the user-code version, it does not sacrifice speed. RoberM already suggested a solution. Ángel, your turn now. ;-)

Dieter

Sorry, I don't have time to do a 100 run. I don't doubt you results since 5 runs is below the minimum of 30 to begin to approach normal distribution.

Namir