- Hi Sam!

Here follows the factorization of 11,236. It began as an ggnfs attempt, but sadly ggnfs isn't quite ready yet to handle a factorization of this scale effortlessly yet. Thus the relations collected was converted into CWI format, and the factorization thus stands as a CWI GNFS C136 factorization.

Credit goes jointly to Tom Cage and Jes Hansen

--- Factorizatiob of 11,236+ ---

Factorization completed after 1791.57 seconds, at Fri Apr 1 11:11:46 2005

Original number had 136 digits:

1988785224338206441736697227355084839346337697529189335828033172544919474764496481130237288855098694100947699948914857840989982990335121

Probable prime factor 1 has 77 digits:

45826489049864977797692632840139991865685587433427696948622219827630008015737

Probable prime factor 2 has 59 digits:

43398158261134913555879300403231730887907132613176534633433

---

Cheers,

Jes

PS: Sam, I've lost my log files. Could you please tell me what numbers from the Cunningham list you have me credited for? - Hi Jes,

I was thinking about asking you if you could send me the relations

on DVD, but it occurs to me that I don't even have a decent enough

machine to do the debugging. The one powerhouse machine that I had

hasn't been the same since it got a new mobo, and my laptop doesn't

have enough RAM. But my suspicion is that GGNFS failed for the same

reason that it has failed on large factorizations for others - this

ll_catFields() re-alloc problem (which, if gone unnoticed, will

probably/possibly cause a matsolve failure, or even a sqrt failure).

Do you happen to have any idea what went wrong?

Cheers,

Chris

--- In ggnfs@yahoogroups.com, jesh@m... wrote:

> Hi Sam!

>

> Here follows the factorization of 11,236. It began as an ggnfs

attempt, but sadly ggnfs isn't quite ready yet to handle a

factorization of this scale effortlessly yet. Thus the relations

collected was converted into CWI format, and the factorization thus

stands as a CWI GNFS C136 factorization.

>

> Credit goes jointly to Tom Cage and Jes Hansen

>

> --- Factorizatiob of 11,236+ ---

>

> Factorization completed after 1791.57 seconds, at Fri Apr 1

11:11:46 2005

> Original number had 136 digits:

>

1988785224338206441736697227355084839346337697529189335828033172544919474764496481130237288855098694100947699948914857840989982990335121

> Probable prime factor 1 has 77 digits:

>

45826489049864977797692632840139991865685587433427696948622219827630008015737

> Probable prime factor 2 has 59 digits:

> 43398158261134913555879300403231730887907132613176534633433

>

> ---

> Cheers,

> Jes

>

> PS: Sam, I've lost my log files. Could you please tell me what

numbers from the Cunningham list you have me credited for? > I was thinking about asking you if you could send me the relations

I've kept the relations because this is an interesting case. Anybody who

> on DVD, but it occurs to me that I don't even have a decent enough

> machine to do the debugging.

want to have a look at the raw relations can just email me with a snailmail

adress, and I will send the relations on some CD's (or DVD's, if that's what

the sender wants) tomorrow. As an alternative I am willing to provide the

files by FTP.

> ll_catFields() re-alloc problem (which, if gone unnoticed, will

I tried the 0.75.1 version complied with the MinGW binaries, but that also

> probably/possibly cause a matsolve failure, or even a sqrt failure).

> Do you happen to have any idea what went wrong?

failed (same error again).

Since I'm neither fluent in C, or able to program in C, I can only suggest

things to people with programming skills:

My gut feeling is that something is very wrong with the way ggnfs handles

the collected relations. The CWI suite had no trouble in factoring 11,236+

with the collected relations. (Which is exactly the ggnfs relations

collected by Tom Cage and I)

I have a dual boot system, so when I ran into the memory trouble with

Windows I tried to compile the programs under Linux.

Compiling the programs under Linux is much easier than under MinGW :-)

When I ran procrels on the collected relations, procrels told me that I

needed to sieve for more relations. Since I knew I needed only 23,4M

relations and I had 25,5M relations, this was where I decided ggnfs was

wrong.

This was the point where I decided to switch to CWI.

--

Cheers,

Jes

>> Original number had 136 digits:

>>

> 1988785224338206441736697227355084839346337697529189335828033172544919474764496481130237288855098694100947699948914857840989982990335121

>> Probable prime factor 1 has 77 digits:

>>

> 45826489049864977797692632840139991865685587433427696948622219827630008015737

>> Probable prime factor 2 has 59 digits:

>> 43398158261134913555879300403231730887907132613176534633433