Hi all,
some of us may have an old Pentium-I or similar system standing around.
Those systems could be set to good use in our mersenne prime search.
They could get factoring or double-checking assignments.
But what if you just want to do first-time tests?
Newcomers may not want to wait many months to get just one result.
Maybe at first they just want to discover a prime that makes it to the top
100.
So they could go for a prime of, say, formula k*2^n+1.
They could run NewPGen to screen a range of numbers of that formula for
small factors.
Then run PRP to test each number in turn for probable primality.
If it finds a probable prime, the Proth program could do the final and
definitive test on that number. It all works really well.
Once they would discover such a big prime -- which usually doesn't take very
long to do -- they may have developped a taste for it, and decide to join in
and help us in our search for a new #1.
But what about the reliability of the result of each test in PRP?
Prime95 reports errors such as 'ILLEGAL SUMOUT' and roundoff errors if they
occurr during the Lucas-Lehmer test.
I'm wondering if PRP is about equally good at detecting errors during a
calculation?
Maybe it would be useful to show a 64-bit residue after completion of the
probable prime test? Re-running the same test would then tell you if the
number is indeed composite.
For now, it looks to me like GIMPS is the most reliable way of looking for
primes.
Does anyone on the list have a view on this?
Robert van der Peijl
Eindhoven, The Netherlands
_________________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
