Firstly, I'm using Phi(n) to represent the cyclotomic polynomial whose index is n, assuming n is non-squarefree, and using p,q,r,... for primes.

It's well known that Phi(p) only have the coefficient 1.

It's also well known that Phi(pq) have coefficients bounded in absolute value by 1.

So here're the puzzles:

1a) What can be said about bounds on coefficients of Phi(pqr)?

b) What's the assymptotic growth?

c) For what pqr does Phi(pqr) have the largest coefficients (that you can find)?

2) As 1, but what if you restrict p,q,r to be simultaniously == 1 mod 4, or == 3 mod 4? Is there a significant difference between the two subsets? (Inspired by Gosper)

3a) Can you categorise p,q,r such that Phi(pqr) only has coefficients in {-1,0,1}?

b) Can you find a bigger pqr with that property than anyone else?

Enjoy!

Phil

--

() ASCII ribbon campaign () Hopeless ribbon campaign

/\ against HTML mail /\ against gratuitous bloodshed

[stolen with permission from Daniel B. Cristofani]