Use Pepin's Tests for proving primality of Mersenne numbers ?

[primus]

Originally Posted by T.Rex

But yes, if we could find some probabilist test for Mersenne numbers that would be faster than LLT, that would help sorting good candidate exponents before running the LLT.

Tony

http://math.stackexchange.com/q/1721415

[ewmayer]

Originally Posted by primus

http://math.stackexchange.com/q/1721415

Please stop spewing nonsense - PARI/GP 'implementations' are meaningless in terms of estimating computational speed. "How many large-integer modmuls are needed and does the modulus permit a zero-padded DWT version of such?" is all that matters. Given an (asymptotically) identical modmul count and DWT-friendliness, there is no earthly reason to prefer the probabilistically-prime algorithm over the real deal.