Talk:Find largest left truncatable prime in a given base: Difference between revisions
Talk:Find largest left truncatable prime in a given base (view source)
Revision as of 13:17, 30 November 2012
, 11 years ago→Hint for base 10, 12, 14 etc
(→Hint for base 10, 12, 14 etc: Some of my notes/thoughts) |
|||
Line 10:
: It's fascinating that when I compare the number of primes at different stages using different levels of reliability testing, the difference with one Miller-Rabin round is only about 1% larger than the count with 5, and often rather less than that. This means that as the number of candidates narrows back down again as the prime size increases, the bad candidates seem to be being removed. Fascinating, that. Also, if you've not got a good MR implementation, you really ''need'' a decent <code>modpow()</code> function/operator; it makes a gigantic difference. (Now, to find a faster computer…) –[[User:Dkf|Donal Fellows]] 09:09, 8 October 2012 (UTC)
:: Yes, why is this task so easy? [http://www.chalcedon.demon.co.uk/rgep/rcam.html Richard G.E. Pinch] records his personal research on strong psudoprimes here. Analysis of the data presented there indicates that the Largest Left Truncatable Prime algorithm is a bad generator of strong pseudoprimes, orders of magnitude worse than a random number generator. I have added a task [http://rosettacode.org/wiki/Carmichael_3_strong_pseudoprimes,_or_Miller_Rabin%27s_nemesis here] which introduces the subject to Rosetta Code.--[[User:Nigel Galloway|Nigel Galloway]] 13:17, 30 November 2012 (UTC)
==Number of left truncatable primes in a given base==
| |||