Thinking out loud about getting under 20M unfactored exponents - Page 14

VBCurtis · 2020-05-14, 05:38

Excellent, thank you for confirming my estimates about P-1 with some prior P-1 done.

ECM does seem a less efficient method than re-doing P-1 for candidates with prior B1 of 120k-150k. I'll run ECM anyway for a while and see if my estimates about factor frequency are reasonable; I imagine I'll end up P-1'ing the ones that stand to benefit most after I play with ECM.

lycorn · 2020-05-16, 09:40

I´m back at work in low ranges.
Just started 16M, 69->70. 16.9 is currently underway.

On another note, the overall < 20M subproject is now getting a huge push from the folks at SRBase / BOINC. They are sweeping through high ranges at an amazing rate, finding an impressive number of new factors. Let´s see how long this will last.

VBCurtis · 2020-05-20, 16:22

In the 4.0M range I've chosen to work in, P-1 seems about twice as fast per expected factor as ECM. I have ECM running on an old laptop anyway, with P-1 on a newer machine.

I found my first two factors with P-1 for this effort, 76 and 81 bits. Rather satisfying! For the sake of this effort, too bad they were factors for the same exponent...

When I saw a 48-digit factor in the results file, I was like "awww, yea! I bet that's a record!" (for this size) Alas, just a pair of normal-sized factors. Exponent is 4067587.

petrw1 · 2020-06-03, 17:31

A thousand days have passed since my first assignment in this sub-project.
In those thousand days I:
- Found 1,569 factors
- Completed over 3,000,000 GhzDays of work (mostly thanks to my 2080Ti)
- Got 50M range to less than 10 ranges with over 1,999 factors remaining
- Got 40M to under 200,000 factors remaining

But this is a mere fraction of the total factors found and ranges cleared by the collective.

Wouldn't it be cool if when I get to 1,999 days I can declare that the under-2,000 project is complete!

Thanks to all.

masser · 2020-06-05, 06:33

Quote:

Originally Posted by petrw1

- Got 40M to under 200,000 factors remaining

lycorn · 2020-06-05, 12:22

Also, taking the figures in This post as a reference, and noting that at the time of typing my current post the number of unfactored exponents below 1G is 21 029 537, we see that in a bit more than 2 years we cleared just over 500,000 of them. Not bad, maybe another 5 years to achieve the goal that makes the subject of this thread.

ATH · 2020-06-11, 23:04

Today we fell below 21M unfactored exponents: 20,999,392 as of 22:00 UTC (6 pm EDT).

Edit:
It happened at 19:00 UTC:

18:00 UTC (2 pm EDT): 21,000,058
19:00 UTC (3 pm EDT): 20,999,407

petrw1 · 2020-06-12, 00:33

Quote:

Originally Posted by ATH

Today we fell below 21M unfactored exponents: 20,999,392 as of 22:00 UTC (6 pm EDT).

Edit:
It happened at 19:00 UTC:

18:00 UTC (2 pm EDT): 21,000,058
19:00 UTC (3 pm EDT): 20,999,407

Yes, SRBase is knocking off the factors like a madman...he swept through all the remaining 8 million or so 70-71 bit assignments in a matter of a couple months

petrw1 · 2020-06-12, 00:34

Quote:

Originally Posted by ATH

18:00 UTC (2 pm EDT): 21,000,058
19:00 UTC (3 pm EDT): 20,999,407

By the way where do you see this?

ATH · 2020-06-12, 01:34

My script that generates this page: unfactored.html

I ran it on the hourly backups I have of https://www.mersenne.org/primenet

petrw1 · 2020-07-10, 01:48

In just under 3 years.

65,710 TF for 642 Factors with a 0.98% success rate
31,195 P1 for 972 Factors with a 3.12% success rate
3,095 ECM for 29 Factors with a 0.94% success rate

2020-06-03, 17:31	#147
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 2³·5²·23 Posts	1,000 Days A thousand days have passed since my first assignment in this sub-project. In those thousand days I: - Found 1,569 factors - Completed over 3,000,000 GhzDays of work (mostly thanks to my 2080Ti) - Got 50M range to less than 10 ranges with over 1,999 factors remaining - Got 40M to under 200,000 factors remaining But this is a mere fraction of the total factors found and ranges cleared by the collective. Wouldn't it be cool if when I get to 1,999 days I can declare that the under-2,000 project is complete! Thanks to all.

2020-06-11, 23:04	#150
ATH Einyen Dec 2003 Denmark 6002₈ Posts	Today we fell below 21M unfactored exponents: 20,999,392 as of 22:00 UTC (6 pm EDT). Edit: It happened at 19:00 UTC: 18:00 UTC (2 pm EDT): 21,000,058 19:00 UTC (3 pm EDT): 20,999,407 Last fiddled with by ATH on 2020-06-11 at 23:08

2020-07-10, 01:48	#154
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 2³·5²·23 Posts	I've completed 100,000 assignments for this project In just under 3 years. 65,710 TF for 642 Factors with a 0.98% success rate 31,195 P1 for 972 Factors with a 3.12% success rate 3,095 ECM for 29 Factors with a 0.94% success rate

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Thinking of Joining GPU to 72	jschwar313	GPU to 72	3	2016-01-31 00:50
Thinking about lasieve5	Batalov	Factoring	6	2011-12-27 22:40
Thinking about buying a panda	jasong	jasong	1	2008-11-11 09:43
Loud thinking on irregular primes	devarajkandadai	Math	4	2007-07-25 03:01
Question on unfactored numbers...	WraithX	GMP-ECM	1	2006-03-19 22:16

2020-05-14, 05:38	#144
VBCurtis "Curtis" Feb 2005 Riverside, CA 5·947 Posts	Excellent, thank you for confirming my estimates about P-1 with some prior P-1 done. ECM does seem a less efficient method than re-doing P-1 for candidates with prior B1 of 120k-150k. I'll run ECM anyway for a while and see if my estimates about factor frequency are reasonable; I imagine I'll end up P-1'ing the ones that stand to benefit most after I play with ECM.

2020-05-16, 09:40	#145
lycorn Sep 2002 Oeiras, Portugal 2653₈ Posts	I´m back at work in low ranges. Just started 16M, 69->70. 16.9 is currently underway. On another note, the overall < 20M subproject is now getting a huge push from the folks at SRBase / BOINC. They are sweeping through high ranges at an amazing rate, finding an impressive number of new factors. Let´s see how long this will last.

2020-05-20, 16:22	#146
VBCurtis "Curtis" Feb 2005 Riverside, CA 5×947 Posts	In the 4.0M range I've chosen to work in, P-1 seems about twice as fast per expected factor as ECM. I have ECM running on an old laptop anyway, with P-1 on a newer machine. I found my first two factors with P-1 for this effort, 76 and 81 bits. Rather satisfying! For the sake of this effort, too bad they were factors for the same exponent... When I saw a 48-digit factor in the results file, I was like "awww, yea! I bet that's a record!" (for this size) Alas, just a pair of normal-sized factors. Exponent is 4067587.

2020-06-05, 12:22	#149
lycorn Sep 2002 Oeiras, Portugal 10110101011₂ Posts	Also, taking the figures in This post as a reference, and noting that at the time of typing my current post the number of unfactored exponents below 1G is 21 029 537, we see that in a bit more than 2 years we cleared just over 500,000 of them. Not bad, maybe another 5 years to achieve the goal that makes the subject of this thread.

2020-06-12, 01:34	#153
ATH Einyen Dec 2003 Denmark 110000000010₂ Posts	My script that generates this page: unfactored.html I ran it on the hourly backups I have of https://www.mersenne.org/primenet