140+ digits which is better

[marthamm]

would ecmnet or GT3+QSieve both with 20+ ppl helping, be
best on 'very'large numbers

[xilman]

Originally Posted by marthamm

would ecmnet or GT3+QSieve both with 20+ ppl helping, be
best on 'very'large numbers

Better for what?

That is, what do you know ahead of time about your number? From your mention of ecmnet it sounds a fair bet that you want to factor some integer or integers.

I confess that I do not know what you mean by "QT3+QSieve" and, to be honest, can't be bothered to go searching. If you were to explain in more detail, possibly with references to other descriptions, I'd take a look but life is just too short otherwise.

In particular, it would be helpful to know what you mean by "very large". There's a fair chance it means > 140 digits (deduced from your choice of title) but do you mean 200 digits? Two thousand digits? Two million digits? Even bigger?


Paul

[BotXXX]

It is also a question what kind of number is it? Is it a special number, part of some formula or perhaps a RSA number? The latter one consists of two primes that are equal in size. And trying ecm on such a number is not wise. Ofcourse you might be 'lucky' but you have a much much higher chance to win the big lottery :)

[marthamm]

Well it is an rsa key,so i guess i will try my luck at QS.
BTW GTS/QS quadratic sieve useing a GT3 container

[xilman]

Originally Posted by marthamm

Well it is an rsa key,so i guess i will try my luck at QS.
BTW GTS/QS quadratic sieve useing a GT3 container

Fair enough, though I must warn you that the world record factorization by QS is one which has "only" 135 digits. If you succeed, you'll have taken the record by a substantial margin.

Performing the same factorization by GNFS is likely to take at most one tenth of the amount of computation and probably rather less, depending on the relative efficiencies of implementation.


Paul