Lucas-Lehmer test

[science_man_88]

Originally Posted by eumayer

and what's soo cool about it is that my code can only return 8 mersenne primes, 
where as this one can return moreee because it is only finding the correct p, which is a lot 
smaller than m so my computer can handle it.! if that is all correct! that is awesome

Mathgirl makes me confused

Which 8 Mersenne primes can your computer return only? The first 8 ones?

2²-1, 2³-1, 2⁵-1, 2⁷-1, 2¹³-1, 2¹⁷-1, 2¹⁹-1, 2³¹-1 ?????

If you really find the code to calculate the correct p for new-record Mersenne primes, then you just made GIMPS redundant

But I am pretty much sure u didn´t !

I think what's meant is that when you use the mersenne primes as exponents you would end up with 2^(2^31-1)-1 so the next would exponent tried would overflow a 32 bit register with that mistake where as now She can check all exponents up to the limit of 32 bits instead of the double mersenne numbers that could be prime or not.

[Mathsgirl]

well if you say
for p in range(1, 1000):
LucasLehmertest(p)
# the test I showed everyone earlier
it returns all the correct Mersenne primes, with corresponding primes in the range(1, 1000).
Whereas the code I wrote originally could only return the first 8 Mersenne primes.
My original code looks like this
for p in primes:
# I had created a list of prime numbers
mersenneprimes = [m for m in primes if (log(m + 1))/log(2) in primes]

so my code is checking values that are much, much larger than the LLT is checking. Now my computer can find about the first 20 mersenne primes with the LLT (maybe more, my computer is still running it now)