Let's Optimize P-1 for low exponents. TL;DR in post #1. More in posts 60 and 61. - Page 12

petrw1 · 2022-04-29, 20:46

Quote:

Originally Posted by Uncwilly

I believe that Jocelyn is male and identifies as such.
Jocelyn is a male name in the Francophone world.

Sorry Jocelyn; no offence intended.

DrobinsonPE · 2022-04-30, 04:08

Quote:

Originally Posted by petrw1

Jocelyn has been dabbling in the 3Million ranges doing low bound ECM and P-1 for many years now.
Unfortunately for this project her bounds are much too low ... they may warrant redoing at some point to more reasonable bounds.
If you want to message her I think her id is: jocelynl1204

mikr had been systematically TF'ing all ranges down as low as 4M to 71 bits.
Only recently have I seen him doing P-1....his bounds are a little better.
I'm not sure how to message mikr

Thanks, when I picked the 3.8M range I did not see that someone was already working there. The pool of exponents is big and there is no reason for me to jump in to a range someone else has already been working. I will try to PM both, let them know of this thread and that I will be moving on to another range shortly.

Can you put me down for 4.0M, I will move there next week.

jocelynl1204 · 2022-04-30, 07:21

Quote:

Originally Posted by DrobinsonPE

Thanks, when I picked the 3.8M range I did not see that someone was already working there. The pool of exponents is big and there is no reason for me to jump in to a range someone else has already been working. I will try to PM both, let them know of this thread and that I will be moving on to another range shortly.

Can you put me down for 4.0M, I will move there next week.

In the past couple decades, I have been mainly working on factored numbers in the range 3M. In recent months, I added the unfactored ones to my list. But since there is a common effort in the ranges below 30M for unfactored ones. I will leave the field open for you all in that range. There are plenty of factored ones to keep me happy for another decade; I hope!

lycorn · 2022-04-30, 08:11

Quote:

Originally Posted by kruoli

On mersenneforum.org, there seems to be no such user. Which user name is it? Not mikr.

His forum name is Miszka.

petrw1 · 2022-05-03, 17:06

I officially started on this project April 16; however unofficially it was mid-January when I decided my next project would be this one and I started using the prescribed bounds to complete my last TWOK ranges.

Some of you started working here in March. So if no one objects I will arbitrarily choose March 1 as day 0.

Also, I don't know of any way to easily determine what method was used to find a factor (TF, ECM, P-1, P+1).
And I only know how to easily report newly factored exponents; not total factors found including extra factors for already-factored exponents.

I'd like to be able to include a count of exponents remaining to P-1 but I'm having trouble determining the "guidelines".
This is where mind is at (whenever it is not wandering off somewhere else):
1. I use mersenne.ca to list for every exponent of interest (under 20? 30? 40M?)
--a. Current bounds and TF level
--b. Proposed new bounds
2. Then I need to determine the current and proposed P-1 success percent.
--a. I have the source for prob.php but I'd prefer a much simpler approximation that I could make into a formula.
--b. I'd be content if it was even within 0.5% or less given my proposed target is in the 3% range.
--c. I'd like to believe it could be approximated with a "quadratic??" taking into account: Exponent, TF-Level, B1 and B2.
--d. It just occurred to me that with all the ECM done in the low ranges the calculated P-1 success percent will NOT be very accurate unless it is also taken into account.
3. Propose what success differential is worth pursuing and count those; maybe something around 3%???
--a. I observe that some of you are methodically P-1'ing every exponent if your ranges; I have no problem with this...it's your rig.

So until someone more knowledged enlightens me I will start with this to report total new factors with each update:
PS: The ToDo is arbitrarily Proposed/Current B1 * P/C B2 > 100 ... and I currently only have data up to 20M for this.

Code:

			New	01-May
Range	Exps.	ToDo	Factors	Factored
0M	78498	14789	25	63365
1M	70435	17031	80	53207
2M	67883	18170	43	49567
3M	66330	11897	97	47239
4M	65367	16359	71	45511
5M	64336	15730	41	44551
6M	63799	14054	9	43853
7M	63129	17791	77	43223
8M	62712	15967	107	42831
9M	62090	10970	93	42256
10M	61938	17912	46	42085
11M	61543	8116	0	41670
12M	61192	6885	0	41231
13M	60825	19091	162	40885
14M	60627	17796	265	40660
15M	60426	17252	385	40460
16M	60184	19049	255	40222
17M	60053	18625	117	40077
18M	59683	18965	184	39711
19M	59557	16941	108	39589
20M	59336	0	1	39365
21M	59318	0	0	39341
22M	58960	0	1	39030
23M	58901	0	40	38978
24M	58805	0	0	38828
25M	58600	0	1	38658
26M	58538	0	0	38562
27M	58365	0	0	38397
28M	58246	0	0	38265
29M	58183	0	0	38216
30M	58120	0	0	38149
31M	57836	0	0	37879
32M	57852	0	0	37943
33M	57712	0	0	37924
34M	57396	0	0	37430
35M	57487	0	0	37524
36M	57361	0	0	37398
37M	57343	0	10	37390
38M	57436	0	5	37474
39M	57252	0	0	37304
40M	57102	0	57	37184
41M	56864	0	0	36882
42M	56915	0	35	37018
43M	56849	0	57	36966
44M	56776	0	99	37002
45M	56893	0	0	37003
46M	56640	0	0	36744
47M	56451	0	0	36593
48M	56387	0	0	36471
49M	56603	0	0	36774
TOTALS			2471

DrobinsonPE · 2022-05-04, 02:08

@petrw1

I started on 03-22-2022 so you missed a few exponents in the 4.2M range

4.20M complete with 12 factors found
4.21M complete with 15 factors found
4.22M complete with 17 factors found
4.23M complete with 7 factors found
4.24M @ 37% with 7 factors found so far.
current total 58

petrw1 · 2022-05-04, 02:34

Quote:

Originally Posted by DrobinsonPE

@petrw1

I started on 03-22-2022 so you missed a few exponents in the 4.2M range

4.20M complete with 12 factors found
4.21M complete with 15 factors found
4.22M complete with 17 factors found
4.23M complete with 7 factors found
4.24M @ 37% with 7 factors found so far.
current total 58

Ok; I backed up so day 0 is March 1.

DrobinsonPE · 2022-05-04, 03:22

Quote:

Originally Posted by petrw1

Ok; I backed up so day 0 is March 1.

You didn't have to do that. This is your game you set the rules. I was just bragging.

petrw1 · 2022-05-04, 04:31

Quote:

Originally Posted by DrobinsonPE

You didn't have to do that. This is your game you set the rules. I was just bragging.

No problem... LOL
Keep up the good work

shatterdpixel · 2022-05-13, 05:26

I'll get to work on 8.2 - 8.3
Currently using an 10400f with about 4ghz, and 8 GB of ram.

Edit: Appears someones already hit those, will keep searching for an untaken assignment.
Second Edit: I'll just work on the ones that have been missed.

petrw1 · 2022-05-13, 06:36

Quote:

Originally Posted by shatterdpixel

I'll get to work on 8.2 - 8.3
Currently using an 10400f with about 4ghz, and 8 GB of ram.

Edit: Appears someones already hit those, will keep searching for an untaken assignment.
Second Edit: I'll just work on the ones that have been missed.

Welcome aboard

2022-05-03, 17:06	#126
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 2×3²×17² Posts	My first attempt at a status update. I officially started on this project April 16; however unofficially it was mid-January when I decided my next project would be this one and I started using the prescribed bounds to complete my last TWOK ranges. Some of you started working here in March. So if no one objects I will arbitrarily choose March 1 as day 0. Also, I don't know of any way to easily determine what method was used to find a factor (TF, ECM, P-1, P+1). And I only know how to easily report newly factored exponents; not total factors found including extra factors for already-factored exponents. I'd like to be able to include a count of exponents remaining to P-1 but I'm having trouble determining the "guidelines". This is where mind is at (whenever it is not wandering off somewhere else): 1. I use mersenne.ca to list for every exponent of interest (under 20? 30? 40M?) --a. Current bounds and TF level --b. Proposed new bounds 2. Then I need to determine the current and proposed P-1 success percent. --a. I have the source for prob.php but I'd prefer a much simpler approximation that I could make into a formula. --b. I'd be content if it was even within 0.5% or less given my proposed target is in the 3% range. --c. I'd like to believe it could be approximated with a "quadratic??" taking into account: Exponent, TF-Level, B1 and B2. --d. It just occurred to me that with all the ECM done in the low ranges the calculated P-1 success percent will NOT be very accurate unless it is also taken into account. 3. Propose what success differential is worth pursuing and count those; maybe something around 3%??? --a. I observe that some of you are methodically P-1'ing every exponent if your ranges; I have no problem with this...it's your rig. So until someone more knowledged enlightens me I will start with this to report total new factors with each update: PS: The ToDo is arbitrarily Proposed/Current B1 * P/C B2 > 100 ... and I currently only have data up to 20M for this. Code: New 01-May Range Exps. ToDo Factors Factored 0M 78498 14789 25 63365 1M 70435 17031 80 53207 2M 67883 18170 43 49567 3M 66330 11897 97 47239 4M 65367 16359 71 45511 5M 64336 15730 41 44551 6M 63799 14054 9 43853 7M 63129 17791 77 43223 8M 62712 15967 107 42831 9M 62090 10970 93 42256 10M 61938 17912 46 42085 11M 61543 8116 0 41670 12M 61192 6885 0 41231 13M 60825 19091 162 40885 14M 60627 17796 265 40660 15M 60426 17252 385 40460 16M 60184 19049 255 40222 17M 60053 18625 117 40077 18M 59683 18965 184 39711 19M 59557 16941 108 39589 20M 59336 0 1 39365 21M 59318 0 0 39341 22M 58960 0 1 39030 23M 58901 0 40 38978 24M 58805 0 0 38828 25M 58600 0 1 38658 26M 58538 0 0 38562 27M 58365 0 0 38397 28M 58246 0 0 38265 29M 58183 0 0 38216 30M 58120 0 0 38149 31M 57836 0 0 37879 32M 57852 0 0 37943 33M 57712 0 0 37924 34M 57396 0 0 37430 35M 57487 0 0 37524 36M 57361 0 0 37398 37M 57343 0 10 37390 38M 57436 0 5 37474 39M 57252 0 0 37304 40M 57102 0 57 37184 41M 56864 0 0 36882 42M 56915 0 35 37018 43M 56849 0 57 36966 44M 56776 0 99 37002 45M 56893 0 0 37003 46M 56640 0 0 36744 47M 56451 0 0 36593 48M 56387 0 0 36471 49M 56603 0 0 36774 TOTALS 2471 Last fiddled with by petrw1 on 2022-05-04 at 02:37

2022-05-13, 05:26	#131
shatterdpixel Apr 2022 Oregon, USA 1100₂ Posts	I'll get to work on 8.2 - 8.3 Currently using an 10400f with about 4ghz, and 8 GB of ram. Edit: Appears someones already hit those, will keep searching for an untaken assignment. Second Edit: I'll just work on the ones that have been missed. Last fiddled with by shatterdpixel on 2022-05-13 at 05:31

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to optimize the sieving stage of QS?	Ilya Gazman	Factoring	6	2020-08-26 22:03
Placeholder: When is it legal to torrent BBC tv stuff?	kladner	Lounge	3	2018-10-01 20:32
Future project direction and server needs synopsis	gd_barnes	No Prime Left Behind	6	2008-02-29 01:09
Unreserving exponents(these exponents haven't been done)	jasong	Marin's Mersenne-aries	7	2006-12-22 21:59
A distributed-computing project to optimize GIMPS FFT? Genetic algorithms	GP2	Software	10	2003-12-09 20:41

2022-05-04, 02:08	#127
DrobinsonPE Aug 2020 3·47 Posts	@petrw1 I started on 03-22-2022 so you missed a few exponents in the 4.2M range 4.20M complete with 12 factors found 4.21M complete with 15 factors found 4.22M complete with 17 factors found 4.23M complete with 7 factors found 4.24M @ 37% with 7 factors found so far. current total 58