Let's finish DC of old "dubious" claims of primes

kriesel · 2021-12-24, 20:43

Of the roughly 90 "dubious" claims exponents, I have the following as not yet having a factor found, matching DC or TC, or PRP with proof & successful cert.

Code:

PRP=1,2,86225219,-1,76,0
PRP=1,2,87232913,-1,76,0
PRP=1,2,89023807,-1,76,0
PRP=1,2,96532433,-1,76,0
PRP=1,2,97970567,-1,76,0
PRP=1,2,103404713,-1,76,0
PRP=1,2,108575167,-1,77,0
PRP=1,2,110000017,-1,78,0
PRP=1,2,122628553,-1,78,0
PRP=1,2,137362691,-1,80,0
PRP=1,2,231695603,-1,79,0
PRP=1,2,276036017,-1,80,0
PRP=1,2,285754177,-1,80,0
PRP=1,2,315690521,-1,80,0
PRP=1,2,333333223,-1,81,0
PRP=1,2,369452123,-1,81,0
PRP=1,2,852348659,-1,86,0

Please reserve any you take on. PRP with GEC and proof generation would be preferable, since it is much more reliable and efficient than the other choices.

mathwiz · 2021-12-24, 21:35

Why do these exponents deserve special attention, just because someone made a claim about them? (in other words: don't feed the trolls).

Uncwilly · 2021-12-24, 23:15

Those from the "Guess Mxx" threads should usually be taken with 10g NaCl crystals.

tuckerkao · 2021-12-25, 01:40

Quote:

Originally Posted by Uncwilly

Those from the "Guess Mxx" threads should usually be taken with 10g NaCl crystals.

How about raise the trial factoring depth 2~3 bits higher than the recommended levels, maybe double the P-1 bounds.

kriesel · 2021-12-25, 01:52

Quote:

Originally Posted by Uncwilly

Those from the "Guess Mxx" threads

Overlap is rare.
M852348659 appeared elsewhere and was classified as a dubious claim first, before someone entered it in the predict m52 thread, so it is not counted again in the predict m52 total.
I found no other overlap between my dubious claim DC list in post 1 and the predict-Mxx lists on my server share (M45, M49, M50, M51, M52).
Predict Mn guesses are not counted again if reused in Mo (o>n) guesses, in the counts I occasionally update.

Quote:

with 10g NaCl crystals.

If usual table use size, that is a LOT of grains of salt. If softener salt, still quite a few.

tuckerkao · 2021-12-25, 02:11

Quote:

Originally Posted by mathwiz

Why do these exponents deserve special attention, just because someone made a claim about them?

If it's like a horse race and all users have gotten only 3 active guesses. Unless they on-purposely enter the bad guesses(such as has a known factor, ridiculously large), the final answers to their declarations should be worked on further if Kriesel grants them. The only oversized exponent seems to be M852348659, other entries are valid and looks okay.

A horse can only be declared lost if the result is a certified composite. In addition, there are barely around 90 of them, not 9,000.

Maybe UncWilly should add this rule to the Mxx thread, exponent guesses must not be greater than 4x the size of the exponent of the current largest known Mersenne Prime or 3x the size of the exponent of the current wavefront milestone, whichever one is larger, but it's up to UncWilly to decide.

A lot of complains have been generated because someone places down a large guess in the M800M+ or M900M+ ranges which will take around a year or 2 to test the exponent, there can be an easy fix to this. Is this what you've meant the trolls?

retina · 2021-12-25, 03:07

There is no obligation to test exponents the others have suggested. So there are no grounds for complaints IMO.

If you want to test them then go ahead. Else leave it to the person suggesting to test their own exponents.

If you don't like the their suggestion then just ignore it and carry on with whatever are your own preferred exponents.

Zhangrc · 2021-12-25, 10:50

Quote:

Originally Posted by tuckerkao

How about raise the trial factoring depth 2~3 bits higher than the recommended levels, maybe double the P-1 bounds.

By finding a factor to prove that they are really wrong?

tuckerkao · 2021-12-25, 11:22

M89023807 which is the 3rd exponent on Kriesel's list was already certified from the PRP branch although LL wasn't verified yet.

Quote:

Originally Posted by Zhangrc

By finding a factor to prove that they are really wrong?

That's more of ViliamF's way because he has better graphic cards.

kriesel · 2021-12-25, 12:54

I'm running PRP=1,2,86225219,-1,76,0 on gpuowl on a Radeon VII, which should take ~16 hours. Can't register it because of the old LLDC reservation lingering for years.

kriesel · 2021-12-27, 23:11

Most of these I did a first test on, so the server won't issue DC to me.

Quote:

Originally Posted by kriesel

Of the roughly 90 "dubious" claims exponents, I have the following as not yet having a factor found, matching DC or TC, or PRP with proof & successful cert.
Please reserve any you take on. PRP with GEC and proof generation would be preferable, since it is much more reliable and efficient than the other choices.

Code:

PRP=1,2,87232913,-1,76,0
PRP=1,2,96532433,-1,76,0 I have reserved this one
PRP=1,2,97970567,-1,76,0
PRP=1,2,103404713,-1,76,0
PRP=1,2,108575167,-1,77,0
PRP=1,2,110000017,-1,78,0
PRP=1,2,122628553,-1,78,0 I have reserved this one
PRP=1,2,231695603,-1,79,0
PRP=1,2,276036017,-1,80,0
PRP=1,2,285754177,-1,80,0
PRP=1,2,315690521,-1,80,0
PRP=1,2,333333223,-1,81,0 I have reserved this one
PRP=1,2,369452123,-1,81,0
PRP=1,2,852348659,-1,86,0

2021-12-24, 20:43	#1
kriesel "TF79LL86GIMPS96gpu17" Mar 2017 US midwest 1CC6₁₆ Posts	Let's finish DC of old "dubious" claims of primes Of the roughly 90 "dubious" claims exponents, I have the following as not yet having a factor found, matching DC or TC, or PRP with proof & successful cert. Code: PRP=1,2,86225219,-1,76,0 PRP=1,2,87232913,-1,76,0 PRP=1,2,89023807,-1,76,0 PRP=1,2,96532433,-1,76,0 PRP=1,2,97970567,-1,76,0 PRP=1,2,103404713,-1,76,0 PRP=1,2,108575167,-1,77,0 PRP=1,2,110000017,-1,78,0 PRP=1,2,122628553,-1,78,0 PRP=1,2,137362691,-1,80,0 PRP=1,2,231695603,-1,79,0 PRP=1,2,276036017,-1,80,0 PRP=1,2,285754177,-1,80,0 PRP=1,2,315690521,-1,80,0 PRP=1,2,333333223,-1,81,0 PRP=1,2,369452123,-1,81,0 PRP=1,2,852348659,-1,86,0 Please reserve any you take on. PRP with GEC and proof generation would be preferable, since it is much more reliable and efficient than the other choices. Last fiddled with by kriesel on 2021-12-24 at 20:43

2021-12-25, 12:54	#10
kriesel "TF79LL86GIMPS96gpu17" Mar 2017 US midwest 16306₈ Posts	I'm running PRP=1,2,86225219,-1,76,0 on gpuowl on a Radeon VII, which should take ~16 hours. Can't register it because of the old LLDC reservation lingering for years. Last fiddled with by kriesel on 2021-12-25 at 12:56

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
"Mixed Sierpinski conjecture base 5" proven!! (if probable primes allowed)	sweety439	sweety439	1	2020-12-21 14:08
Is "mung" or "munged" a negative word in a moral sense?	Uncwilly	Lounge	15	2020-04-14 18:35
Aouessare-El Haddouchi-Essaaidi "test": "if Mp has no factor, it is prime!"	wildrabbitt	Miscellaneous Math	11	2015-03-06 08:17
Palindrome primes (a.k.a. Elementary S03E03 "Just a Regular Irregular")	Batalov	And now for something completely different	12	2014-11-16 19:03
Would Minimizing "iterations between results file" may reveal "is not prime" earlier?	nitai1999	Software	7	2004-08-26 18:12

2021-12-24, 21:35	#2
mathwiz Mar 2019 2·163 Posts	Why do these exponents deserve special attention, just because someone made a claim about them? (in other words: don't feed the trolls).

2021-12-24, 23:15	#3
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 10891₁₀ Posts	Those from the "Guess Mxx" threads should usually be taken with 10g NaCl crystals.

2021-12-25, 03:07	#7
retina Undefined "The unspeakable one" Jun 2006 My evil lair 1A19₁₆ Posts	There is no obligation to test exponents the others have suggested. So there are no grounds for complaints IMO. If you want to test them then go ahead. Else leave it to the person suggesting to test their own exponents. If you don't like the their suggestion then just ignore it and carry on with whatever are your own preferred exponents.