[QUOTE=bsquared;283163]I've seen this behavior myself and found it annoying, but I haven't had time to do anything about it yet. I agree that it should run more curves at the previous bound before increasing the bound, up to a limit (probably the gmp-ecm "optimal" curve count for that bound).

Other planned changes:
* ECM progression to follow gmp-ecm optimal bounds when searching for factors progressively 5 digits larger (i.e., ... 1M, 3M, 11M, 43M, ...)
* plan = deep will start at the 1M ECM mark, instead of the 250k mark as it does now.[/QUOTE]

What would yafu do if ECM finds e.g. a c3x composite factor with plan=deep? Something like TF / rho / ECM 15@5e4 / QS on the c3x?

[QUOTE=Andi47;283174]What would yafu do if ECM finds e.g. a c3x composite factor with plan=deep? Something like TF / rho / ECM 15@5e4 / QS on the c3x?[/QUOTE]

Yes, it would recurse on the composite factor starting with TF. With a c3x, I doubt it would do any ecm, and would probably go straight to QS after rho (or possibly pm1).

You probably already know it, but the idea behind plan=deep is to use it on inputs for which you already know there are no small factors.

[QUOTE=bsquared;283163]I've seen this behavior myself and found it annoying, but I haven't had time to do anything about it yet. I agree that it should run more curves at the previous bound before increasing the bound, up to a limit (probably the gmp-ecm "optimal" curve count for that bound).

Other planned changes:
* ECM progression to follow gmp-ecm optimal bounds when searching for factors progressively 5 digits larger (i.e., ... 1M, 3M, 11M, 43M, ...)
* plan = deep will start at the 1M ECM mark, instead of the 250k mark as it does now.[/QUOTE]
Maybe there should be a limit on how high bounds ECM will ever use.

Like in the case of the C118, I don't think it should go any higher than 3e6. Perhaps not even that high.

Stupid question. The function "siqs" is faster using the linux or the windows client?

[QUOTE=pinhodecarlos;283496]Stupid question. The function "siqs" is faster using the linux or the windows client?[/QUOTE]
Should be the same I believe. The nfs siever is faster on linux if that's what you are getting confused with.

Stupid question...Is there a command I can use to tell YAFU to factor a number in a .txt document? I tried >factor(wtd.txt) and >factor "wtd.txt", but to no avail...

[QUOTE=c10ck3r;284659]Is there a command I can use to tell YAFU to factor a number in a .txt document?[/QUOTE]

Make a text-file say "n.txt" with content like:
[code]
factor(15716973377858501123247575740273676440789062959869051082006740376262967297236329270279)
[/code]

and call yafu with:
yafu <n.txt

How exactly do I "call yafu"? Is this just a command line entry? Or do I type it in yafu?

[QUOTE=c10ck3r;284672]How exactly do I "call yafu"? Is this just a command line entry? Or do I type it in yafu?[/QUOTE]

In WIN start a DOS-box (Start -> Execute -> cmd) and type "yafu <n.txt".

[QUOTE=c10ck3r;284672]How exactly do I "call yafu"? Is this just a command line entry? Or do I type it in yafu?[/QUOTE]

Command line entry. E.g. instead of starting yafu by double clicking it or by running "yafu", run "yafu <n.txt" or make a shortcut pointing to "yafu <n.txt".

[QUOTE=c10ck3r;284659]Stupid question...Is there a command I can use to tell YAFU to factor a number in a .txt document? I tried >factor(wtd.txt) and >factor "wtd.txt", but to no avail...[/QUOTE]

from the command prompt,

[CENTER]yafu "factor(@)" -batchfile a.txt > out.txt[/CENTER]