Leading the way to the Future
 13 Aug 2010 07:35:05 am by Weregoose (Details) – Had I done this during high school when I first attempted it, I would have considered it my magnum opus... This is valid for numbers up to 9,080,191. More testing (and maybe further optimizing) after I get some sleep. Ans→C For(T,0,1 C-1→S Repeat fPart(S 1→X 42T+31→Y .5S→S Ans→B While Ans If fPart(.5Ans round(CfPart(XY/C),0→X round(CfPart(Y[font=verdana]²/C),0→Y int(.5B→B End If X+1=C or X=1 [font=times new roman]π→S End AnsfPart(.5C) or max(C={2,31,73 If Ans End Ans
 13 Aug 2010 11:03:05 am by thornahawk Looks great! I've been trying to figure out how to do Solovay-Strassen and the Lucas pseudoprime tests properly myself. Speaking of which, there's also an implementation of Miller-Rabin in http://www.jjj.de/fxt/fxtpage.html (get the PDF of the book there); you might be able to pick up something useful. thornahawk
 13 Aug 2010 07:17:08 pm by Weregoose (PDF) I tried the {2,299417} pair to get an upper limit of 19,471,033; unfortunately, rounding matters chewed it up and spat out wrong results, so I'll have to keep the bases small. {2,7,61} looks to be fairly popular on the interwebs – I'll modify the code above and test it out. Though, if I'll be having the calculator take up time with more passes, then I'll just go all-out and get that 1012 I've been aiming for. I'll just go with these three bases first. [EDIT] Hmm... Drat. For(T,0,1 → For(T,0,2 42T+31→Y → 2+int(.28[font=times new roman]×[font=arial]√T5→Y C={2,31,73 → C={2,7,61 Those should have cut it. 4,759,123,129 is erroneously marked as composite. (Not even 10,135,421 works...) Moving forward.