mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Cunningham Tables

Reply
 
Thread Tools
Old 2016-09-15, 23:01   #210
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

1000100101112 Posts
Default

And 2,1019+ is done. Bruce had a fighting chance at this one!

frmky is online now   Reply With Quote
Old 2016-09-18, 14:33   #211
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

23518 Posts
Default

Quote:
Originally Posted by frmky View Post
And 2,1019+ is done. Bruce had a fighting chance at this one!
Do you mean a fighting chance with ECM curves?
Bruce Dodson's limit is being 74 digits, Sam Wagstaff's limit is being 79 digits, Ryan Propper's limit is being 83 digits.
77 digits is clearly not at Bruce Dodson's limit at all!
My limit - good question! 39 digits - but at IIT Madras department computing facility spent most of time cracking Cunningham tables low hanging fruits with SNFS rather than instead of ECM curves. Not even GNFS at all.
Right now that I am trying it out ECM curves up on over at following four numbers (28191+1)/3, (219937+1)/3, (2110503+1)/3, (2524287+1)/3 with in higher bounds on a four core micro processor.

And then or: Go advanced.
Missing out with in some thing?

There has been appearing to be no new GMP-ECM top 10 records results in the year 2016 tables since at least Wednesday 20 April 2016. Really? Why?
Has been it is being the case that Paul Zimmermann may be that he went out Missing In Action up? Or it is being too difficult to find out one single up right now also?
Or people are not being focusing up on over with in ECM curves with in right now also?
And then with in ECM curves with in:
And then or too:
What is Joppe Bos' team being doing right now?
At their own will, they could finish off with in entire Cunningham tables, Fibonacci numbers, Lucas Numbers, Homogeneous Cunningham Numbers and other twisted additive or multiplicative groups like these things.
May be it is being case that NFS@Home should also to consider out factoring with in Coppersmith SNFS - Factorization Factory variant too!
NFS@Home may be that it could be case also to consider out factoring with in (21123+1)/3, (3661-1)/2, (7379-1)/6, (7379+1)/8 too!
No non-primitive, non-trivial factors are being known out for these Cunningham table numbers up any way at all!
Not too hard to solve factoring, consider tackling also these numbers out are being up!
No further more world record factorizations champions are being possibly available out any way for a while up!
Why?
And then or also too with in ECM curves with in:
Now that the Cunningham tables have been extended to a homogeneous, i.e. uniform level such that the number of digits with in the last number entry is being balanced enough!
Not like before where ever that the base 3 + and base 3 - tables were being lagging behind.
To calculate with extension limit × log10base with in order to check out with in.

1300 × log102 = 391.34.
850 × log103 = 405.55.
550 × log105 = 384.43.
500 × log106 = 389.08.
450 × log107 = 380.29.
400 × log1010 = 400.00.
350 × log1011 = 364.49.
350 × log1012 = 377.71.

Base 6 exponent index power limit = 500.
Base 7 exponent index power limit = 450.
Base 8 exponent index power limit = 433 ⅓.
Base 9 exponent index power limit = 425.
Base 10 exponent index power limit = 400.

For Fibonacci numbers, Lucas Numbers tables which share many similar properties with the Cunningham tables,
Algebraic and derived factors, Aurifeuillian Factors for Lucas numbers whose exponent is a multiple of 5,
Every prime factor of Fibonacci numbers and Lucas numbers always of form 2kp±1, every prime factor of Lucas numbers are always of form 1 (mod 10) or 9 (mod 10). Why?
Consider with in exponent index, power growth rate α = (1+√5)/2, and β = (1-√5)/2
such that Fibonaccin = (αnn)/(α-β), and Lucasn = (αnn)
α, β are being roots of the equation x2 - x - 1 = 0.
Feasible Fibonacci numbers, Lucas Numbers tables limit = 2000 such that as like for the posting it that in to Mersenne Wiki web site pages:

2000 × log10α = 417.98.

Not like before where ever that the imbalance does occur heterogeneously do ununiformly.

1200 × log102 = 361.24.
600 × log103 = 286.27.
450 × log105 = 314.54.
400 × log106 = 311.26.
400 × log107 = 338.04.
400 × log1010 = 400.00.
300 × log1011 = 312.42.
300 × log1012 = 323.75.

Base 6 exponent index power limit = 400.
Base 7 exponent index power limit = 400.
Base 8 exponent index power limit = 400.
Base 9 exponent index power limit = 300.
Base 10 exponent index power limit = 400.

Feasible Fibonacci numbers, Lucas Numbers tables limit = 1500 such that as like for the posting it that in to Mersenne Wiki web site pages:

1500 × log10α = 313.48.

Last fiddled with by Raman on 2016-09-18 at 15:31
Raman is offline   Reply With Quote
Old 2016-09-19, 13:09   #212
bdodson
 
bdodson's Avatar
 
Jun 2005
lehigh.edu

210 Posts
Default

Quote:
Originally Posted by Raman View Post
Do you mean a fighting chance with ECM curves?
Bruce Dodson's limit is being 74 digits, Sam Wagstaff's limit is being 79 digits, Ryan Propper's limit is being 83 digits.
77 digits is clearly not at Bruce Dodson's limit at all! ...
According to this description, no one would ever improve on their
best current effort; set as a limit. Any factor p70 or above is still
an achievement (fewer than 25, total, so far); but I consider any
p7x to be a plausible ecm target, and include smaller p8x's.
Rather than setting a limit, those p79 and p83 are targets to
be aimed for. There is a difference in the factors p75-and-up
(a total of seven, exclusively due to Sam and Ryan), they're
being found with limits p65/p70, even p75, only possible on
large memory workstations. For contrast, my p74 was found
on a commodity pc, an i7, on one of our public labs.

Ryan has a post on the p75-limit run; he says that he
submitted 15K curves, one of which hit the p83 (as I recall).
I'm up to 70K curves with B1 = 3e9, p70-optimal on our
old cluster, with c. 100K curves expected to find a p70.
Most of those on wanted Cunninghams that Greg has
shown to have no chance what-so-ever of an ecm factor;
say p90-or-up. So p77 may not be a likely target, but
there is a relatively plausible fighting chance. -Bruce
bdodson is offline   Reply With Quote
Old 2016-09-19, 15:08   #213
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3·419 Posts
Default

I meant till now. My previous (prose) post - first para graph (essay) passage - was being (especially exactly essentially) a joke. Every one can be to improving up on their own ECM curve record efforts. Please do not take it out seriously - that does up!

Okay, could NFS@Home consider finishing off with in the entire Cunningham Tables numbers by using Coppersmith SNFS?
Fibonacci numbers, Lucas Numbers, Homogeneous Cunningham Numbers and other twisted additive or multiplicative groups like these things.
This Factorization Factory variant is being not usable for GNFS candidates numbers, such that they need to be tackled separately!

I don't think that Texas State HPC which will make out any progress up after having been holding on to this number for over six years of time period frame duration - time times know known!
Some body else could claim - do tackle with in this GNFS candidate number.

Ryan Propper does have got with in his own resources, so been he can surely tackle with in 2,1067+ 3,671- 3,671+ apart from computation of Aliquot Sequence 4788 / 314718.
Bruce Dodson can also surely tackle too 6,448+ besides of ECM curves - Lehigh University Condor Cluster.
Chen + Fang + Reich + Eichhorn + Voznyy + Beschorner finished off with in 10,343- a while back!
Raman is offline   Reply With Quote
Old 2016-09-19, 23:59   #214
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

3·733 Posts
Default

Quote:
Originally Posted by Raman View Post
I don't think that Texas State HPC which will make out any progress up after having been holding on to this number for over six years of time period frame duration - time times know known!
Some body else could claim - do tackle with in this GNFS candidate number.
No. They report still making progress on it. It is theirs. There are plenty of others to work on.
frmky is online now   Reply With Quote
Old 2016-09-21, 15:15   #215
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3×419 Posts
Default

Quote:
Originally Posted by Raman
Well, let me know very well about this thing around - does variably.
think stuff style item step idea

Nothing but noting that - that ever which ever a way a way - ever to - up on over on to to to.

presently currently a way a way by right now itself - at this very moment variably - ever?
Any one has got with in an ETA (Estimated Time of Arrival) been for Texas State HPC 2,2158M c193?
Coppersmith SNFS - Factorization Factory variant is being suboptimal for smaller numbers - size of numbers length that ever which ever NFS@Home is being doing it that with in right now,

New speed up algorithms will emerge out as we do tackle with in harder composite number candidates!

Last fiddled with by Raman on 2016-09-21 at 16:12
Raman is offline   Reply With Quote
Old 2016-09-22, 11:45   #216
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3×419 Posts
Default

Every one except just a very few other people have stayed away right now from these types of factoring efforts effectively and efficiently so far.

I could have just done 10,590M c213 and 3,605+ c185 including those from extension tables at resources of IIT Madras department computing facility but I later decided to give up because:

10,590M c213 some one else would be able to sieve better and faster than me by using three large prime numbers.
3,605+ c185 from those extension tables had not been extensively done ECM curves running / executing as like the regular Cunningham tables so waste of a computational effort if a small factor lands and results.

As like the regular Cunningham tables, now those from extension tables have now just been extensively done ECM curves running / executing?

Every one except just a very few other people have stayed away right now from these types of factoring efforts effectively and efficiently so far.

Last fiddled with by Raman on 2016-09-22 at 11:54
Raman is offline   Reply With Quote
Old 2016-09-22, 14:39   #217
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

256616 Posts
Default

Quote:
Originally Posted by Raman View Post
I could have just done 10,590M c213 and 3,605+ c185 ...

Every one except just a very few other people have stayed away right now from these types of factoring efforts effectively and efficiently so far.
Assumes facts not in evidence.
Batalov is offline   Reply With Quote
Old 2016-10-01, 11:23   #218
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3×419 Posts
Default

Quote:
Originally Posted by Batalov View Post
Assumes facts not in evidence.
But got mentally disturbed or upset and when an administrator deleted 120 million range of special-q sieved, relations accumulated for 2,2334L on Friday 4 February 2011.
May be he had been thinking that folder directories like 2_2334M, 2_1930M and 2_2334L were automated scripts or viruses.
And I had to sieve again.
Nevertheless I had extra time to sieve latter numbers like 10,590M and 3,605+ after wards the former numbers like 2,2334M, 2,1930M and 2,2334L were sent out for compute cluster linear algebra.
At least sent out to linear algebra for some one else.
6,343- linear algebra I had been sending it out off to Tom Womack in year of March 2009.
2,985- finished off by Joppe Bos by using SNFS quickly. (May be that is Coppersmith Factorization Factory variant?) It is out off of list then!
Is good to better needed to consider with for better verifying off with for factors again.
As is prime factorization is not leading out to unique results away with for, different runs executing away will produce with for different results!
Raman is offline   Reply With Quote
Old 2016-10-01, 16:36   #219
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2×4,787 Posts
Default

Quote:
Originally Posted by Raman View Post
Is good to better needed to consider with for better verifying off with for factors again.
Anyone can verify the complete set of tables of Cunningham factorizations in under 1 second. Then they can verify them again on different hardware in under 1 second. Then they can verify them again with ten different software implementations in a few seconds.
And for those with cyberphobia, there is always a ream of paper and a pen (and an abacus). Take any factor that you suspect to be "wrong" and divide it using school arithmetic; it is not that hard!
Quote:
Originally Posted by Raman View Post
As is prime factorization is not leading out to unique results away with for, different runs executing away will produce with for different results!
Say what? With thoughts like these, you better join the (solitary) conversation in this thread.
Batalov is offline   Reply With Quote
Old 2016-10-22, 22:52   #220
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

219910 Posts
Default

2,1079+ is done.

frmky is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
5+ table garo Cunningham Tables 100 2021-01-04 22:36
7+ table garo Cunningham Tables 86 2021-01-04 22:35
6+ table garo Cunningham Tables 80 2021-01-04 22:33
5- table garo Cunningham Tables 82 2020-03-15 21:47
6- table garo Cunningham Tables 41 2016-08-04 04:24

All times are UTC. The time now is 22:09.


Wed Oct 27 22:09:14 UTC 2021 up 96 days, 16:38, 0 users, load averages: 1.05, 1.08, 1.07

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.