Bases 2 & 4 reservations/statuses/primes

[Jean Penné]

Riesel Base 2 odd n's to-day :

k = 133977 is up to n = 524429 no prime, so I unreserve this k...
k = 86613 is up to n = 348807 no prime, continuing up to 512K

Reserving now k = 148323 up to n = 512K

Regards,
Jean

[Jean Penné]

As I reported on an other thread here, k = 86613 is eliminated, because :

86613*2^356967-1 is prime! (code L591) ; now 15 k's are remaining.

I am presently testing two k's :

k = 145257 is up to n = 304853
k = 148323 is up to n = 331647

The 16 presieved files are now sieved up to p = 1.018 tera, so I updated them on my personal site :

http://jpenne.free.fr/ConjRus/

I am continuing the sieving only on the 15 remaining k's, so it becomes faster.

Karsten, do you intend to update the data on the Liskovets-Gallot page of Riesel Prime Search soon? There are pretty many changes to report!
Thank you by advance, and Best Regards,
Jean

[kar_bon]

found this 2 days ago:
(thanks Jean for correction)
198726*2^268878-1 = 198726*4^134439-1 = 99363*2^268879-1

and today

260934*2^273436-1 = 260934*4^136718-1 = 130467*2^273437-1

so there're 13 k's remaining!

by now all these 13 k's tested upto n=274,000 -> odd n=137,000!

[kar_bon]

i updated the page with all information from last posts!

thanks Jean for the sieving files for Riesel odd n, so another 2 k's kicked out (see above)!

[kar_bon]

198726*2^268878-1 = 198726*4^134439-1 = 99363*2^268879-1

and

260934*2^273436-1 = 260934*4^136718-1 = 130467*2^273437-1

[Jean Penné]

Many congrats, Gary and Karsten, for these last three primes, there are very nice results, because 1 k is eliminated for Sierpinski base 2 odd n's and 2 k's are eliminated for Riesel base 2 odd n's.

Moreover, the big sievings I started for these two sub-projects will become a lot faster!

Please, Karsten would you credit yourself, (instead of me) for the two primes you discovered!

Best Regards,
Jean

[Jean Penné]

Originally Posted by kar_bon

i updated the page with all information from last posts!

thanks Jean for the sieving files for Riesel odd n, so another 2 k's kicked out (see above)!

Thank you Karsten!
I am also continuing to sieve, on another machine, the 8 remaining files for Sierpinski base 2 even and odd n's (which also cover the complete Sierpinski base 4 project).
As I chose a large range (n = 1 to 16M base 2), the progress is slower, but I am already reaching 70 billions ; I shall make a first distribution when 100 billions reached.
These data will be usable both for testing and double-checking...
Regards,
Jean

[gd_barnes]

Originally Posted by kar_bon

found this 2 days ago:
(thanks Jean for correction)
198726*2^268878-1 = 198726*4^134439-1 = 99363*2^268879-1

and today

260934*2^273436-1 = 260934*4^136718-1 = 130467*2^273437-1

so there're 13 k's remaining!

by now all these 13 k's tested upto n=274,000 -> odd n=137,000!

Based on this, I am reserving for Karsten all remaining k's that are not already reserved by Jean and showing them as tested to n=274K.

odd n=137000 ??? I hope not.

Do you mean base 4 n=137000?


This brings up something that has caused a little confusion in the past:

Can I request that we stick with a specific standard when showing primes or search-ranges? That standard should be whatever base the conjecture is and whatever the k is for the conjecture is how we should state testing limits or primes. I don't understand the redundancy of showing the even-k/even-n and the base 4 primes here when we're proving odd-k/odd-n for base 2.

The only exception might be if the k is remaining on both base 2 and base 4, which is usually not the case, especially for Riesel base 2 odd-n with a high conjecture.

Here, just stating 99363*2^268879-1 and 130467*2^273437-1 as prime with a testing limit of n=274K base 2 on all k's should be all that is needed.


Thanks,
Gary

[Jean Penné]

Originally Posted by gd_barnes

Based on this, I am reserving for Karsten all remaining k's that are not already reserved by Jean and showing them as tested to n=274K.

odd n=137000 ??? I hope not.

Do you mean base 4 n=137000?


This brings up something that has caused a little confusion in the past:

Can I request that we stick with a specific standard when showing primes or search-ranges? That standard should be whatever base the conjecture is and whatever the k is for the conjecture is how we should state testing limits or primes. I don't understand the redundancy of showing the even-k/even-n and the base 4 primes here when we're proving odd-k/odd-n for base 2.

The only exception might be if the k is remaining on both base 2 and base 4, which is usually not the case, especially for Riesel base 2 odd-n with a high conjecture.

Here, just stating 99363*2^268879-1 and 130467*2^273437-1 as prime with a testing limit of n=274K base 2 on all k's should be all that is needed.


Thanks,
Gary

Gary,
I am now totally confused about the reservations for this sub-project :
Do you mean that now, I may not reserve any more k's for this sub-project?
So, I should be forced to keep the three k's I am working on for ever, and I think it is not a good strategy! (that is what I did for two years on Sierpinski base 4, and I don't wish to continue this way...) On the contrary, I think that when the exponents become so large (and the chance to find a prime so small) it is better to split the presieved files so the provers can take small ranges...
I need more specifications about this...
Regards,
Jean

[gd_barnes]

No, I'm not talking about reservations here. You can reserve any k you want in any base. You can reserve as many k's or bases that you can test! I'm only referring to POSTING of primes and ranges completed.

I'm saying that when you and Karsten post a prime or a range completed, please just post it in the format of the Conjectured k, base, and exponent and leave out forms of the primes and ranges that don't pertain the specific conjecture(s).

Here's an example on Karsten's post:

198726*2^268878-1 = 198726*4^134439-1 = 99363*2^268879-1

and today

260934*2^273436-1 = 260934*4^136718-1 = 130467*2^273437-1

so there're 13 k's remaining!

by now all these 13 k's tested upto n=274,000 -> odd n=137,000!

My question is why are we showing 3 different forms of the same prime? These k's are not remaining for anything else except riesel odd-n. So...what I'm asking is that we keep the prime and ranges posts short and simple with:

99363*2^268879-1 is prime

and today
130467*2^273437-1 is prime

so there're 13 k's remaining!

all these 13 k's now tested up to n=274,000

In this case, we are not testing even k's, we're not testing even n's, and we're not testing base 4. We're just testing Riesel base 2 odd-n.

So there's no reason to state the even k's, even n's, or base 4 forms of the primes.

Now...if one test covers more than one base or conjecture, THEN it makes sense to state them in more than one way. For instance k=9519 for base 2 even-n and for base 4.

I hope this clarifies the request.

Please know that this isn't a requirement...it's only a suggestion. I'm asking because I've actually posted the incorrect n-value for a prime on my pages when reading one of these previously. I then caught it the next day.

Also, I would never limit anyone from searching anything as long as they are providing regular updates. That's the fastest way to lose people from a project.


Thanks,
Gary

[Jean Penné]

Originally Posted by gd_barnes

No, I'm not talking about reservations here. You can reserve any k you want in any base. You can reserve as many k's or bases that you can test! I'm only referring to POSTING of primes and ranges completed.

I'm saying that when you and Karsten post a prime or a range completed, please just post it in the format of the Conjectured k, base, and exponent and leave out forms of the primes and ranges that don't pertain the specific conjecture(s).

Here's an example on Karsten's post:




My question is why are we showing 3 different forms of the same prime? These k's are not remaining for anything else except riesel odd-n. So...what I'm asking is that we keep the prime and ranges posts short and simple with:




In this case, we are not testing even k's, we're not testing even n's, and we're not testing base 4. We're just testing Riesel base 2 odd-n.

So there's no reason to state the even k's, even n's, or base 4 forms of the primes.

Now...if one test covers more than one base or conjecture, THEN it makes sense to state them in more than one way. For instance k=9519 for base 2 even-n and for base 4.

I hope this clarifies the request.

Please know that this isn't a requirement...it's only a suggestion. I'm asking because I've actually posted the incorrect n-value for a prime on my pages when reading one of these previously. I then caught it the next day.

Also, I would never limit anyone from searching anything as long as they are providing regular updates. That's the fastest way to lose people from a project.


Thanks,
Gary

O.K. I think it is now clear, and I agree with you about how to state the results.
It remains that the reserved ranges for each k might always be well known by all participants in order to avoid duplicate work, and it is not so easy to assert that always...
Jean