10+ table

[Raman]

Originally Posted by Max0526

Sooner or later nfs@home, Kurt Beschorner, ryanp, or someone else will need it:

Msieve v. 1.52 (SVN 958)
R0: -211749473706922200328618506605766270339
R1: 25149014102345040136349927
A0: -64775438425752496430873578692913572727347501664
A1: 53063705532362222569585460244930918034
A2: 34917546543547006314477513393123
A3: -100918039437211814160901
A4: -1955199543988376
A5: 784320
skew 150000000.00, size 2.436e-019, alpha -6.959, combined = 7.240e-015

I think that NFS@Home has not reserved with 10,325+ c197
because in my own created up Mersenne Wiki web site page,
(currently maintained up by using Jyb, Wreck and Shaopu Lin?)
it is being in SNFS Quartics list, not GNFS list.

NFS@Home has been reserving up numbers looking up at the my own constructed up
Mersenne Wiki web site page tables at all?

Is it easier by using SNFS Quartic rather than GNFS? It is closer off.
But I think that NFS@Home is right now in form of doing with GNFS candidates
instead of SNFS Quartic, or Quintic or Sextic.

That along with 2,2562L c196 from formerly added up extension tables is being
in SNFS Quartics list, not GNFS list.

NFS@Home has been getting reserved up with all GNFS candidate numbers up to c208 including
the old one 2,1285- c218.

Numbers like 3,619- c207 and 3,671- c205 (right now reserved by using Ryan Propper)
might have been got easier by using SNFS Sextic, not by using GNFS at all!

I will be updating with Mersenne Wiki tables sooner right now, along with Cunningham
number tables, Fibonacci and Lucas number tables, Homogeneous Cunningham number tables
in their own beautiful coloured derived prime factors format! It would be good enough
to have with Java Script enabled or PHP enabled in to Mersenne Wiki web site page to
automate with submission process of prime factors at all!

May be that many an other people like it up? Or not?

[Raman]

Originally Posted by Raman

Rather than = instead of?
Size = length? In MATLAB programming code / GNU Octave language - size returns with in a two dimensional matrix always exactly only certainly, but not only, let not alone, length returns with in a one dimensional vector always exactly only certainly, namely - not - not!
Cause = effect?

If no one does with these following for a while,
Then they could be becoming with Smaller but needed GNFS candidates!

With the state of art, some of these candidates
it may be that easier by using GNFS rather than
by using SNFS quartic or quintic or sextic
polynomial candidates.

Up on the my own created up and constructed up
Mersenne Wiki web site page, it might be
that needing up as like some adjusting up type
of thing at all, any way some how every day.

2,2158M c193 2,2562L c196 10,325+ c197
12,299+ c202 3,748+ c204 3,671- c205
Of the form
3,619- c207 7,356+ c207 2,2330L c207
2,1037+ c209 3,643+ c210 2,2330M c210
Skipped up, snipped up
2,2210M c211 3,715- c211 12,305+ c212
7,833M c212 11,293+ c212 2,1100+ c213
Cf. What does <snip>,
12,274+ c213 2,1240+ c216 3,725- c217
2,1165+ c217 3,820+ c218 11,334+ c219
<skip>, Cf. mean out?
3,631+ c219 2,2126M c219 7,395- c220
2,1255- c220 7,413- c220 3,632+ c220
Have been got with
7,403+ c220 3,815- c220 2,2494M c221
11,649L c221 7,395+ c221 7,425- c221

3,718+ c222 ...

It might be that good enough up to having been
got with sorting up by using size or length
rather than instead of difficulty of candidates
SNFS quartic or quintic or sextic polynomial
candidates.

10 Most Wanted numbers.
24 More Wanted numbers.
How many Smaller but needed numbers?

2,2158M c193 (Reserved up by using Texas State HPC - Don Hazlewood)
3,671- c205 (Reserved up by using Ryan Propper)

Originally Posted by Raman

• Samuel Wagstaff's Cunningham table home page contains 10 most wanted numbers and 24 more wanted numbers. What is being the secret behind this selection of the numbers 10 and 24?

[wreck]

About 10,325+, here is some informations I have sent to Kurt.

2015-08-01

I have explain the step to get combined value of polynomial
through sina email, and I think other two persons maybe interest in it.

For number larger than snfs 220 (again I forgot the accurate number, or we just do not know it), we should use degree 6, if we do not use degree 6, than there will be a penalty.

Some one says for a degree 10, the penalty is 60.
This maybe the reason why Yousuke is still doing 1620L.

For degree 4, I get the formula from you from another person (sorry I forgot his name),
this formula is some thing like,for a degree 4, its real diff value is
(snfs 4 difficulty) * 6 / 4 - (a constant value)
I forgot the accurate constant value is , maybe is 40.
So , for R650,
its difficulty is 260 * 6/4-40=350,
it is more than 300,
to get the accurate value, I test its speed, and found its speed is 18 sec/rel.
about 32 times slower than R423.
While Max use lpba=31, for a number if its lpba increase 1, its rel collect speed will increase 2 time (while needed relations will increase 2 times too).
For snfs number, when the diff increase 10, its speed will slower 2 times, My test speed using lpba=32 for R423 is 0.6 sec/rel, if I use lpba=31, its speed will be 1.2 sec/rel.
So the real speed of test R650 is 16 times than R423.
so R650's real difficulty (equivalent to snfs 6) is
(R423's snfs diff) + 10 * log2(16)
=282 + 10 * 4
=322

2016-10-13

For R650, I think after R2700M relation collect is done , and when that
time NFS@Home does not reserve it, you could reserve it from Wagstaff.
It should done by you and Max, Wenjie, etc with 6 months.

[Max0526]

Found by CADO-NFS, optimized by Msieve:

Msieve v. 1.52 (SVN 958)
R0: -429873155922929313315767352804496526959
R1: 1173407499622534153549
A0: -4123232726098214293264497150572823602694091032000
A1: -15046303477474735468514048031694219326776
A2: 1269995212430296791873190620103689
A3: -2988888945327618127510111
A4: -23844831853933266
A5: -6981480
skew 249134139.39, size 1.372e-019, alpha -8.945, combined = 4.999e-015

[Max0526]

After some more root sieve in CADO-NFS:

Msieve v. 1.52 (SVN 958)
R0: -429873155894842908328972286223207900523
R1: 1173407499622534153549
A0: 921152784545133512672868444928211620544589981840
A1: -39293694975142056979694617415210276238112
A2: -972446599487622683766964276896261
A3: 5311864367775954917911807
A4: 24680367142186866
A5: 6981480
skew 232590778.72, size 1.422e-019, alpha -9.026, combined = 5.111e-015

[kurtb]

I am reserving 10,515,+/c199 for our team.

Team: Max0526, wreck2002, Wenjie Fang, Alfred Eichhorn and I.
Someone wants to join our group (for R515,+)?

Kurt

[frmky]

10,325+ is done.

[Max0526]

@kurtb
After CADO's second approach:

# norm 1.587617e-019 alpha -6.405212 e 4.818e-015 rroots 3
skew: 244424374.97
c0: 61273646333056248631393145049744489856776032064
c1: 10420727777495411411387820457257803823560
c2: 145201440064897386425846123841612
c3: 259088764547269247676230
c4: -3111278570304511
c5: 1252800
Y0: -262207264182244965566501989754094701687
Y1: 35260964440695209765221

Y0: -260505402482337066318147974102921098728
Y1: 7716435693806557117099
c0: -1654261163502007156972422319406119913455210822815
c1: 26384190129329608515063756157530472625990
c2: 203788580791257674994807537648741
c3: -3672705324080165167457858
c4: -4687498736820054
c5: 11648340
skew: 236770061.88
# size 1.304e-019, alpha -7.791, combined = 4.866e-015 rroots = 3

Y0: -429873155923612753953255001219423760068
Y1: 1173407499622534153549
c0: 3684169692376062463639486252058906152472281984050
c1: 16528721029913580823331834978501372839223
c2: -1275169246253808999055456284843666
c3: 2933359798287776393299687
c4: 23824500352969866
c5: 6981480
skew: 248078670.32
# size 1.332e-19, alpha -8.860, combined = 4.905e-15 rroots = 5

Y0: -256801803982980786519466270774361363475
Y1: 46489253566470781807399
c0: 394135635881473880558342455190986592866466927488
c1: -4990753553766026834416226588003267492240
c2: -176630505265024070591230831366460
c3: -5116291946618703658808340
c4: 9192608450724867
c5: 9732240
skew: 236539316.33698
# alpha -8.23, 3 real roots
# MurphyE=5.01099448e-15

[kurtb]

thank you Max.
sieving start: March 16
10,1350M/c197: postprocessing start with TD=130: March 19
Kurt

[pinhodecarlos]

10, 302+ c289 188981422179250214477885038956646476812007525220846625175628245017547495717341304519447280552146559165713534073382085460954497219653965265520569. p146 NFS@Home snfs

http://pastebin.com/LQkg9tjW

[wreck]

10,1350M is factored by Team "Bo Chen, Wenjie Fang, Maksym Voznyy, Alfred Eichhorn and Kurt Beschorner".
c197 = p83 * p114
The polynomial is found by Maksym using CADO, with a negative c5,
which trigger a problem when doing the post-processing's -nc3 stage.
After change each coefficients of algebraic and rational polynomial,
and then doing the -nc3, the factors pop out successfully.
I guess, perhaps it is possible to do this step automatically
after give some modification to the msieve code when A5 is negative.

Post-processing log is attached.