[VA] Short & Sweet Math Challenge #24: "2019 Spring Special 5-tier"

[John Keith]

Somehow I had completely missed Tier 2 until I saw Bernd's post #35. Then I thought I had a good program until I saw Albert's reply and realized the uniqueness requirement, so back to the drawing board.

This problem turns out to be a good fit for the 50g and the Prime, both of which have NEXTPRIME and ISPRIME? as built-in functions.

My program also uses the I->BL command plus a couple of other commands from ListExt. I have tried to keep stackrobatics to a minimum in the interest of readability.



%%HP: T(3)A(R)F(.);
\<< I\->R \-> b n
\<< { } 1 1. n
START NEXTPRIME DUP b I\->BL LSUM DUP
IF ISPRIME?
THEN DROP
ELSE ROT SWAP DUP2
IF POS
THEN DROP SWAP
ELSE + OVER + SWAP
END
END
NEXT DROP DUP SIZE 2. / LDIST EVAL
\>>
\>>

Inputs are the base on level 2 and the number of primes to check on level 1. Output are two separate lists, the composites and the primes.

I would classify the size (163 bytes) and speed as reasonable if not exactly prize-winning, and it is sort of cheating as it uses so many pre-existing commands. I shudder to think of writing such a program on a "classic" era machine.

I have checked the first 100000 primes for 7 and 31, which take over 5 minutes each on the emulator, so my results are nowhere near as extensive as Albert's. Still a neat problem, I only wish I had noticed it earlier.