The Original 'Ode To Joy' thread

[edorajh]

I reached the top 300 in Top Factoring Producers Report!

[dswanson]

Topped 3000 exponents LL-tested

Account ID      LL P90*  Exponents  Fact.P90  Exponents  P90 CPU
                CPU yrs  LL Tested  CPU yrs*  w/ Factor  hrs/day
--------------  -------  ---------  --------  ---------  -------
dswanson        1789.47     3001       6.070       90    7238.34

[jinydu]

The new computer my parents got me for University just finished its first LL test today!

[leifbk]

I've got five crunchers online:

13331737 D   65               7.8  27.2  80.2                   07-Nov-04 19:12  bulbasar      701 v19/v20
20705683     67               3.3  10.8  63.8  14-Nov-04 06:53  12-Nov-04 08:53  bombadil     2791 v19/v20
21318937     67    126081    10.2  15.7  75.7  15-Nov-04 06:25  05-Nov-04 09:03  brutalis     2391 v19/v20
22971173     68   5326783    53.3  46.8  62.8  21-Oct-04 09:09  23-Sep-04 07:42  buldrian      938 v19/v20
26415877     68  11669951    16.4   5.0  65.0  09-Nov-04 14:36  30-Oct-04 04:46  balapapa     3072 v19/v20

Whoooay!

[JHansen]

I just factored my first number with SNFS difficulty >200 digits (208 to be exact ):

I factored the 190 digit composite in 3^437+1 as follows:

Factorization completed after 2518.94 seconds, at Thu Nov 25 12:17:41 2004
Original number had 190 digits:
1161329893417327062944988188703363237248711650400988615384263277426412987406938676899346214009353007655969012743101733225846006459497055046698739492930329138167502311231751603961882987702489
Probable prime factor 1 has 95 digits:
68547576903992521115550335292598768465537822909888714468554198504012153521955464650549693339597
Probable prime factor 2 has 50 digits:
30526952192487621416237379827468678387890146149269
Probable prime factor 3 has 45 digits:
554983474084953530805865496135990056978910473

Hoping to complete my first +700 bit factorization within a two months.
-----
Cheers,
Jes

[xilman]

Originally Posted by JHansen

I just factored my first number with SNFS difficulty >200 digits (208 to be exact ):

I factored the 190 digit composite in 3^437+1 as follows:

Factorization completed after 2518.94 seconds, at Thu Nov 25 12:17:41 2004
Original number had 190 digits:
1161329893417327062944988188703363237248711650400988615384263277426412987406938676899346214009353007655969012743101733225846006459497055046698739492930329138167502311231751603961882987702489
Probable prime factor 1 has 95 digits:
68547576903992521115550335292598768465537822909888714468554198504012153521955464650549693339597
Probable prime factor 2 has 50 digits:
30526952192487621416237379827468678387890146149269
Probable prime factor 3 has 45 digits:
554983474084953530805865496135990056978910473

Hoping to complete my first +700 bit factorization within a two months.
-----
Cheers,
Jes

Nice one!

Good luck with the 700-bit factorization. Which one is it, as a matter of interest?


Paul

[JHansen]

Originally Posted by xilman

Nice one!

Good luck with the 700-bit factorization. Which one is it, as a matter of interest?


Paul

I'm doing 2^709+1, a C192.

I hope to finish sieving within a month, and then I expect a month worth of linear algebra.

-----
Cheers,
Jes

[JHansen]

Originally Posted by JHansen

I'm doing 2^709+1, a C192.

I hope to finish sieving within a month, and then I expect a month worth of linear algebra.

Actually, there more to it that that. This is my first factorization, where I'm using 29-bit large primes, so I will be dealing with a set of relations, that is rather huge compared to what I'm used to see when using 28-bit Large Primes.

Any "Good Advice(TM)" on how to cope with relation-sets containing this many relations would be greatly appreciated. How do I keep the memory reqs. down? I have about 2 GB's of RAM to play with.

Would the "-filtmin 10M" for the CWI tools be the option to consider to use on the raw data?

-----
Cheers,
Jes

[xilman]

Originally Posted by JHansen

Any "Good Advice(TM)" on how to cope with relation-sets containing this many relations would be greatly appreciated. How do I keep the memory reqs. down? I have about 2 GB's of RAM to play with.

I'll pass on some of my experience via email. I doubt very much that many others will be interested in command line options to programs they don't have access to.

2G RAM is plenty. I've filtered data consisting of 80M relations with up to 2 large primes, each up to 30 bits on machines with 2G. It does get rather harder, though still possible, to process them on 1G systems.

I've done the post-processing of two different numbers where the large primes ran to 29 bits on a 1.3GHz Athlon with 768M RAM. No real problem, though the matrix for one took 6 days and the other 60 hours. I'd hope that the linear algebra for your number doesn't take as long as a month.

Paul

[patrik]

I just reached 20 CPU-years of trial factoring. Starting out with almost 11 years two months ago (done a long time ago), I put two of my hyper-threaded processors doing both LL tests and trial factoring. And I'm climbing fast in the stats. However my LL speed decreased a bit (but I am still running LL tests faster than half the speed).

Account ID      LL P90*  Exponents  Fact.P90  Exponents  P90 CPU
                CPU yrs  LL Tested  CPU yrs*  w/ Factor  hrs/day
--------------  -------  ---------  --------  ---------  -------
patrik          865.216      766      20.086       71    3613.81

[jinydu]