> When George originally created the list of candidate exponents, he > eliminated tens of millions of composite exponents, and an > infinite number > of negative exponents, non-integer exponents, imaginary > exponents, and > prime exponents above the range of the program. > > Should he get credit for all of these, especially given that > it probably > took him an afternoon of programming and computation > combined, tops, to > create the list? > > By your reasoning he should, since he removed the need for LL > testing all > those exponents...
Ahh, but none of those numbers would have been Mersenne numbers anyway by definition. :) The real work was probably just chugging through the first 53 or 54 bits of factoring. Relatively easy on the old CPU, and probably eliminated quite a bit right off the bat with a minimum of effort. I know it makes no difference to the project if a 53 bit factor is found, or a 65 bit factor. It may have taken longer to find the 65 bit one, but each still saved an entire LL test, so... On the one hand, the person doing the factoring might want to get more credit for factoring the larger #, because it took more CPU time to find. :) At any rate, figuring out how much "points" to award for doing trial factoring, and then if you even find a factor... You probably just gotta balance that in a way to encourage just the right amount of factoring work. :) Once the server (and client) are ready to assign P-1 work, this will probably all change a bit more to encourage some P-1 factoring work ahead of the LL tests. And maybe assigning 64 bit factoring, then P-1, and then 65 (and higher) bit factoring... In that order. (that was the optimum way to do it, right?) :) _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
