"-y"... mmmmm, yummyyyy... I was looking for this option in v10
Thanks!

Not quite a bug, just a slight oddity (from sequence 530082):

[code] 1003 . c104 = 82392635655155865748638869421212849653732674881549532138920170145688316038241281878949155624352834699602 = 2 * 7 * 11 * 149 * 257953 * 464767 * 670048307608279 * 44699166466232525862332791932845611574436301222473745137643713569832446353 :[COLOR="Blue"]Decreased[/COLOR]!
next driver: [COLOR="Blue"]Decreased[/COLOR]![/code]

Shouldn't that read "downdriver!"? :wink:

My understanding: decreased is stronger then the downdriver. The program will always write the stronger (most significant) evolution.

A guide is some combination of factors for which the probability is higher for the next terms of the sequence to increase in size. Not necessarily to increase the number of digits, it can go from 12354 to 23456, staying into 5 digits, but increasing in value. Also, not necessarily to increase at all, some terms could decrease, but in average, they will increase.

A driver is just a very persistent guide. If the probability for the next term is higher to decrease, then is a downdriver. There is not necessarily for a downdriver to decrease the next term. If you have a "lot of small factors", then some terms can still be higher then the previous terms. But the "lot of small factors" come and go, and in average your sequence will decrease, if you have a downdriver. Theoretically, the sequence can decrease without downdriver too, so to be more detailed, we could have (1)"decreased without downdriver", (2)"decreased and downdriver", (3)"increased and downdriver". Yes, (3) is also possible, imagine you have something like 2*5*11*59*67*x, it will raise for some terms, even if it has downdriver, until the small factors are gone, then it will fall.

So, "decreased" here is stronger, it means that you got the downdriver AND the term indeed decreased.

[QUOTE=LaurV;279028]My understanding: decreased is stronger then the downdriver. The program will always write the stronger (most significant) evolution.
[/quote]

not necessarily: the line says "next driver". So if the sequence has acquired the downdriver, it is supposed to say either "Downdriver!", or (not necessarily) "Downdriver, decreased".

[quote]A driver is just a very persistent guide. If the probability for the next term is higher to decrease, then is a downdriver. There is not necessarily for a downdriver to decrease the next term. If you have a "lot of small factors", then some terms can still be higher then the previous terms. But the "lot of small factors" come and go, and in average your sequence will decrease, if you have a downdriver. Theoretically, the sequence can decrease without downdriver too, so to be more detailed, we could have (1)"decreased without downdriver", (2)"decreased and downdriver", (3)"increased and downdriver". Yes, (3) is also possible, imagine you have something like 2*5*11*59*67*x, it will raise for some terms, even if it has downdriver, until the small factors are gone, then it will fall.

So, "decreased" here is stronger, it means that you got the downdriver AND the term indeed decreased.[/QUOTE]

There is only one driver, which *can* decrease the sequence*), and this is 2 (not squared, cubed, etc.) without a 3, so this is *the* downdriver. All odther drivers increase the sequence (or keep it constant if it's the perfect number alone without any other factors, e.g. alq_6 = 2*3). As you stated correctly, there are some guides which may decrease the sequence.

Hence the line says "current driver" or "next driver", I would consider the presence of the down[I]driver[/I] stronger (or as strong as) the message "decreased".

[SIZE="1"]*) Note to the nitpickers: "increasing" resp. "decreasing" a sequence is used here as short term for: the next term will be bigger (resp. smaller) than the current one.[/SIZE]

[QUOTE=Andi47;279039]
There is only one driver, which *can* decrease the sequence*), and this is 2 (not squared, cubed, etc.) without a 3, so this is *the* downdriver. All odther drivers increase the sequence (or keep it constant if it's the perfect number alone without any other factors, e.g. alq_6 = 2*3). [/QUOTE]
I have to disagree here, try starting from 44 (2^2*11) or from 2^3*17, or from 2^4*37, and generally from any 2^k*p, where p is the next prime higher then 2^(k+1)-1, or start from (almost any) odd number. There are MANY situations where the next term can decrease. Most of them are not "persistent" from a term to another (so we can not really call them "downdrivers" or "downguides") but some of them are !!!

[QUOTE=LaurV;279040]I have to disagree here, try starting from 44 (2^2*11) or from 2^3*17, or from 2^4*37, and generally from any 2^k*p, where p is the next prime higher then 2^(k+1)-1, or start from (almost any) odd number. There are MANY situations where the next term can decrease. Most of them are not "persistent" from a term to another (so we can not really call them "downdrivers" or "downguides") but some of them are !!![/QUOTE]

These are not [I]drivers[/I], but [I]guides[/I].

[QUOTE=Andi47;279039]not necessarily: the line says "next driver". So if the sequence has acquired the downdriver, it is supposed to say either "Downdriver!", or (not necessarily) "Downdriver, decreased".



There is only one driver, which *can* decrease the sequence*), and this is 2 (not squared, cubed, etc.) without a 3, so this is *the* downdriver. All odther drivers increase the sequence (or keep it constant if it's the perfect number alone without any other factors, e.g. alq_6 = 2*3). As you stated correctly, there are some guides which may decrease the sequence.

Hence the line says "current driver" or "next driver", I would consider the presence of the down[I]driver[/I] stronger (or as strong as) the message "decreased".

[SIZE=1]*) Note to the nitpickers: "increasing" resp. "decreasing" a sequence is used here as short term for: the next term will be bigger (resp. smaller) than the current one.[/SIZE][/QUOTE]

Yes, all this is true. It's just a side-effect of the way I implemented the driver detection -- a decreasing sequence trumps all. Now that I know more I could probably do it differently, but at the time I just wanted to make sure that I never gave up on a decreasing sequence even if my earlier driver detection wasn't working right.

The "next driver:" line is showing the driver of the term currently being factored; 'decreased' is relative to the just-completed term.

I have found a problem with aliqueit. It seems to be unable to factorize the squares of relatively large numbers. In particular for the sequence from attached file Aliqueit gets stuck on the number

366301321 = 19139^2

At least on my computer it just repeats saying:
[QUOTE]WARNING: couldn't find factor in factor.log
WARNING: couldn't find factor in msieve.log
WARNING: qs failed to find a factor. This really shouldn't happen.
I'll cheerfully loop and try again though...[/QUOTE]

Does anyone have Linux binaries (x86-64)? I'd rather not have to DL GMP and compile it, though I will if necessary... also, I recall reading that it can be used with only YAFU somewhere, but I don't remember where I saw that. Is this true?

[QUOTE=Dubslow;295852]Does anyone have Linux binaries (x86-64)? I'd rather not have to DL GMP and compile it, though I will if necessary... also, I recall reading that it can be used with only YAFU somewhere, but I don't remember where I saw that. Is this true?[/QUOTE]I can't speak for the *nix side of things, but with Windows, you can configure it to use either YAFU or msieve depending on preference.....

Also, since Aliqueit does only light factoring, the 32/64 bitness doesn't really affect the speed that much. Also (though I would have to check this), I think the 32-bit compile should be able to call out to 64-bit tools (YAFU/msieve, GMP-ECM, and the PERL/python scripts for GGNFS) for the heavy lifting.

(OK, I just tried it and the 32-bit Windows version can at least call 64-bit GMP-ECM, so I would have to assume the same holds true for YAFU or msieve too....)

The 'whatsnew' mentions a -y option; I ask because currently I can do everything straight through Yafu, and don't have the scripts, gmp-ecm, etc... setup separately. (Actually, I tried installing gmp-ecm out of repositories and that appeared to work, but I have no idea what version it is or if it works or not).

Edit: I have now compiled GMP and now Aliqueit, so if anyone [i]else[/i] needs Linux blobs, let me know.