mersenneforum.org  

Go Back   mersenneforum.org > Extra Stuff > Miscellaneous Math

Reply
 
Thread Tools
Old 2021-09-06, 10:35   #287
Happy5214
 
Happy5214's Avatar
 
"Alexander"
Nov 2008
The Alamo City

3·11·23 Posts
Default

@tuckerkao If you're really desperate for a factor, a second P-1 test to ridiculously high bounds will probably get you a better return than your repeated P+1 tests with the same bounds and different seeds.
Happy5214 is offline   Reply With Quote
Old 2021-09-07, 01:11   #288
mathwiz
 
Mar 2019

3×67 Posts
Default

Quote:
Originally Posted by tuckerkao View Post
If all exponents from 82589939 to 167999989 are composite, then my dozenal math is correct. If not, then it's wrong.
You have presented no actual math here, just wild speculation and numerology.

You keep rambling about "dozenal math", your "skyscraper", etc. It's all just gibberish. There is no actual math to support the claim that your proposed exponent range is any more likely, or any less likely, to contain a Mersenne prime.

My only hope is that others aren't led down this pointless crusade, and just keep their mprime / Prime95 clients receiving whatever assignments the server hands out.
mathwiz is offline   Reply With Quote
Old 2021-09-11, 03:46   #289
tuckerkao
 
"Tucker Kao"
Jan 2020
Head Base M168202123

1111101102 Posts
Default

Quote:
Originally Posted by thyw View Post
168202123 PRP test (~1144 GHzD) assigned on 2021-08-04. Still not finished, but in the meantime, spent ~3272 GhzD of tf and pm1. Too bad you didn't find a factor, so you could throw out your prp test.
Don't know where you got the idea that all those close PM1 105211111 was efficient and to be followed.
It seems like Nvidia Geforce 3080 Ti(GPU on a friend's PC) is still a little slower than the ideal model I want for the trial factoring. 2^83 to 2^84 of M168202123 takes a little less than 3 full days long.

Nvidia Geforce Lovelace 4090 maybe the ideal model I want for the trial factoring which should be able to cover any available exponent choice on the Mersenne number list, the release date is Oct 2022 according to the link below -
https://wccftech.com/nvidia-geforce-...-launch-rumor/

M168202123 should be finished in around 5 more hours including the proof file hashes. As soon as I get the PRP roof placed on the headquarter skyscraper with a certification, I'll run more P-1 on other M168,***,*23 exponents.

Quote:
Originally Posted by Happy5214 View Post
@tuckerkao If you're really desperate for a factor, a second P-1 test to ridiculously high bounds will probably get you a better return than your repeated P+1 tests with the same bounds and different seeds.
I only want the headquarter skyscraper to have a certain shape on the right side, thus I run the P+1 tests with the same bounds and different seeds on different friends' PCs to check out whether their machines are fast enough for me.

I also trial factored up the 3 closest unfactored neighbors on each side of M168202123 up to 2^77, just in case the traffic get busy someday in the future -
https://www.mersenne.org/report_expo...ll=1&ecmhist=1

By the way, I spent 15 hours to run 1 curve of ECM, so that my PRP test will for sure finish on Sep 11th -
https://www.mersenne.org/report_expo...ll=1&ecmhist=1
Attached Thumbnails
Click image for larger version

Name:	Mersenne 168202123 PRP Progress at 99.38%.png
Views:	26
Size:	54.3 KB
ID:	25645   Click image for larger version

Name:	Mersenne Trial Factoring Progress 09-10-2021.png
Views:	26
Size:	112.0 KB
ID:	25646  

Last fiddled with by tuckerkao on 2021-09-11 at 04:29
tuckerkao is online now   Reply With Quote
Old 2021-09-12, 00:55   #290
tuckerkao
 
"Tucker Kao"
Jan 2020
Head Base M168202123

2×251 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
Don't start that range until your current range is finished and proven unprofitable.
I've been working on M168,***,*23 all the time and will do so continuously.

For some reason, P-1 with B1=1,000,000 and B2 = 40,000,000 for M168,***,*23 only take around 14 hours each to complete instead of 22 hours several months ago. This is really a big help.

The exponents I've worked up recently -
M168199123: NF to 2^77 and NF-PM1
M168217123: NF to 2^77 and NF-PM1
M168243223: NF to 2^77 and NF-PM1

New PRP-Ready Exponents since 08/04/2021: M168020323, M168133123, M168175123, M168207223, M168221023, M168227023, M168922123, M168932123, M168934123


Previous PRP-Ready Exponents: M168314323, M168412723, M168424723, M168479323, M168704323, M168779323

It was really a big grace that Borek Boissy picked up M168174323 and its PRP result will be revealed in around 5~6 more days.

There must be something I didn't work efficiently before Aug 2021, but suddenly got it correct by coincidence since the beginning of this August.

Last fiddled with by tuckerkao on 2021-09-12 at 01:03
tuckerkao is online now   Reply With Quote
Old 2021-09-12, 01:15   #291
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

132×59 Posts
Default

Quote:
Originally Posted by tuckerkao View Post
There must be something I didn't work efficiently before Aug 2021, but suddenly got it correct by coincidence since the beginning of this August.
I understand that coincidence is really important in the numerology domain.
chalsall is offline   Reply With Quote
Old 2021-09-12, 08:39   #292
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

5,813 Posts
Default

From prime95's whatsnew.txt:
Code:
New features in Version 30.4/30.5 of prime95.exe
------------------------------------------------

1) Faster P-1 stage 2.
2) Faster ECM stage 1 and stage 2.
3) Gwnum library overhauled.  Many functions deprecated.  Replaced by more powerful gwmul3.  New functions
   that compute (a+b)*c and (a-b)*c with less memory accesses.  Faster conversion to and from binary.
4) ECM and P-1 can find the best B2 value for the amount of memory prime95 is allowed to use.  For ECM,
   this happens when the worktodo.txt line sets B2=100*B1 which is the default assignment from the PrimeNet
   server.  For P-1, the best B2 is chosen when the worktodo.txt line specifies the trial factoring depth.
   For example, "Pminus1=1,2,20000003,-1,500000,0,70" chooses the best B2 bound for B1=500000 given that
   M20000003 has been trial factored to 2^70.
kriesel is online now   Reply With Quote
Old 2021-09-12, 19:14   #293
kruoli
 
kruoli's Avatar
 
"Oliver"
Sep 2017
Porta Westfalica, DE

23·5·17 Posts
Default

Quote:
Originally Posted by tuckerkao View Post
It seems like Nvidia Geforce 3080 Ti(GPU on a friend's PC) is still a little slower than the ideal model I want for the trial factoring.
We have a saying that goes Lieber den Spatz in der Hand als die Taube auf dem Dach (roughly [I]Better to have a sparrow in one's hand than a pigeon on one's roof[/I]).
kruoli is offline   Reply With Quote
Old 2021-09-12, 23:59   #294
Viliam Furik
 
Viliam Furik's Avatar
 
"Viliam Furík"
Jul 2018
Martin, Slovakia

683 Posts
Default

Quote:
Originally Posted by kruoli View Post
We have a saying that goes Lieber den Spatz in der Hand als die Taube auf dem Dach (roughly [I]Better to have a sparrow in one's hand than a pigeon on one's roof[/I]).
We have literally the same saying here, in Slovakia. I am not sure, whether we got it from German language (this is a rather likely scenario, given the history of Slovakia - during the Hungarian rule period, most of the intellectuals here were from Germany, Austria, and such), or it's from a common source. I'll look into it someday.
Viliam Furik is offline   Reply With Quote
Old 2021-09-15, 02:50   #295
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
"name field"
Jun 2011
Thailand

100110010010012 Posts
Default

We have it the same, except is a crow on the fence, not a pigeon on the roof, and as I guess, it is either of Latin, Slavic, or Turkish (Ottoman) origin.

Last fiddled with by LaurV on 2021-09-15 at 02:51
LaurV is offline   Reply With Quote
Old 2021-09-18, 06:11   #296
tuckerkao
 
"Tucker Kao"
Jan 2020
Head Base M168202123

2×251 Posts
Default

Quote:
Originally Posted by LaurV View Post
Wow!
That is a big move, from 123 to *23, ten times more exponents.

Then, when a prime will at last be found, you'll start screaming "told you so!".
I will only say so if the drought that UncWilly described last all the way into M167999989.

New PRP Ready Exponents since 09/11/2021: M168217123, M168316123, M168612823, M168731923

Last fiddled with by tuckerkao on 2021-09-18 at 06:13
tuckerkao is online now   Reply With Quote
Old 2021-09-21, 00:16   #297
tuckerkao
 
"Tucker Kao"
Jan 2020
Head Base M168202123

2×251 Posts
Default

Quote:
Originally Posted by sweety439 View Post
The number of Mersenne exponents end with E should be less than that of Mersenne exponents end with 1, 5, or 7, this is because for all Sophie Germain primes p == 3 mod 4 (all such primes end with E), Mp is divisible by 2*p+1, thus Mp is composite.
It took me some time to find a dozenal Ɛ ender exponent in the M168M range that can actual survive the larger P-1 factoring and has a close M168,***,321 predecessor neighbor and also shares the same type of numerical combinations that I've been looking for in both the dozenal and the decimal bases.

M168174323 locates too far away from M168173321 and has other unfactorable exponents (NF up to 2^82 and NF-PM1 with B1 = 2,800,000 and B2 = 120,000,000) in between the 2 specified exponents. Having a median factored M168,***,*23 exponent which is M168173323 is another issue.

There are no other unfactorable exponents in between M168620423 and M168620321 which makes me think this dozenal Ɛ ender candidate will be worth a PRP try. M168620423 -> Z4857931Ɛ also dozenally has all odd digits in the sequential orders for the last 6 digits, thus easy to remember.

There still other exponents that may fit this characteristic.

Quote:
Originally Posted by thyw View Post
168202123 PRP test (~1144 GHzD) assigned on 2021-08-04. Still not finished, but in the meantime, spent ~3272 GhzD of tf and pm1. Too bad you didn't find a factor, so you could throw out your prp test.
Don't know where you got the idea that all those close PM1 105211111 was efficent and to be followed.
TF from 2^76 to 2^77 for M168620423: 90.76 GHz days, completed on my GPU in around 7 hours.
P-1 with B1 = 2,800,000 and B2 = 120,000,000 for M168620423: 90.64 GHz days, completed on my CPU in around 1 day and 16 hours.

Quote:
Originally Posted by drkirkby View Post
  1. Working at 168 million is not sensible
  2. Choosing your own B1 and B2 is not sensible.
I manually chose B1 = 2,800,000 and B2 = 120,000,000 for M168620423, but the choice wasn't random, the very same 2 bounds have been recommended for all the M461,***,**7 exponents.

Quote:
Originally Posted by Uncwilly View Post
Remember this post?
Real random things can be quite lumpy.
Patience grasshopper.
I found 8 new factors during Aug 2021, however 0 in Sep 2021 so far.

Last fiddled with by tuckerkao on 2021-09-21 at 00:56
tuckerkao is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Automatic fetch of Trial Factoring work for GPU mfakt* LaurV GPU to 72 81 2020-12-02 05:17
Simple Script to get Trial Factoring Work jfamestad PrimeNet 3 2016-11-06 20:32
Why trial factoring work chopped into chunks? lidocorc PrimeNet 4 2008-11-06 18:48
How does the trial factoring work with 15K*2^n-1 jocelynl 15k Search 0 2003-07-11 14:23
How does trial-factoring work? ThomRuley Software 5 2003-05-30 20:34

All times are UTC. The time now is 03:51.


Thu Oct 28 03:51:36 UTC 2021 up 96 days, 22:20, 0 users, load averages: 1.92, 1.96, 2.02

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.