Riesel base 3 reservations/statuses/primes

[gd_barnes]

Originally Posted by KEP

4223272*3^134986-1 is prime.

This means only 1 k to go

Now a quick question, for Gary:

The primes for Riesel base 3 for k<=500M for n<=500 are you interested in having those or will you rediscover those primes again, once you continue constructing or others continue constructing the proof for Riesel base 3 conjecture?

KEP!

Great find!

Thanks for the reminder on primes n>=500. Yes, please send those to me.

[KEP]

Originally Posted by gd_barnes

Great find!

Thanks for the reminder on primes n>=500. Yes, please send those to me.

OK I'll do. It will save you 20 days of work on a single core

[KEP]

Originally Posted by gd_barnes

Great find!

Thanks for the reminder on primes n>=500. Yes, please send those to me.

Ops... I just saw that you thought I reminded you of the primes where n is greater than 500, when in fact I was asking about your interest in getting the primes where n is less than or equal 500. So are you interested in those primes or will you find them yourself later on?

Sorry for askin again, especially if you made a typeo and in fact was answering my first question

KEP

[gd_barnes]

Originally Posted by KEP

Ops... I just saw that you thought I reminded you of the primes where n is greater than 500, when in fact I was asking about your interest in getting the primes where n is less than or equal 500. So are you interested in those primes or will you find them yourself later on?

Sorry for askin again, especially if you made a typeo and in fact was answering my first question

KEP

Oops. My bad.

I did misread your statement but the answer is the same:

Yes, go ahead and send me the primes for n<=500 if the files are not too big.


Thanks,
Gary

[KEP]

Originally Posted by gd_barnes

Oops. My bad.

I did misread your statement but the answer is the same:

Yes, go ahead and send me the primes for n<=500 if the files are not too big.


Thanks,
Gary

Well it will sum up to about 260MB of data (compressed)... but as soon as you've downloaded the data from your e-mail, you can delete the files, so only temporarily it will require a lot of space

Thanks for your answer, I'll now go ahead and start uploading the e-mails, to the draft folder, and as soon as the entire range is uploaded, I will e-mail all parts around the same time, to avoid misunderstandings or data losses

KEP

[KEP]

Originally Posted by henryzz

i dont know why i said sieving i meant primality proving
once my base 15 effort is finished to n=25k i will do 500000 ks from base 3 to n=25k
i have had a rather large gap between primes almost 1/3 of the range tested

Well then I think I've an answer to your previously question. Running 1M range (500,000 k's) to n<=25K, will take about 12 hours, if you only use OpenPFGW and starts out by doing some PRP testing at first, and eventually verifies the PRPs. So for administrative purpose i would suggest that you at least reserves 10M ranges or maybe 100M (dependent on the amount of cores you tent to put on this effort). I'm considering to launch an attack on a 1G range as soon as my Quad is done with the few important reservations she is working on This should take about 150 days from start to finish on the Quad (Q6600).

Also I may add, sieving is far more efficient from n>1000 (maybe n>2500) since trial division and factoring then starts to be to time demanding. But for the easyness of creating the proof later on, I'm considering to do it this way:

1. PRP test all k's reserved to n<1000
2. Sieve the k's remaining for n>1000 to n<=25000
3. PRP test all k's remaining in sieve file (for at most 1 prime per k)
4. Proof the PRP with n>1000
5. Proof the PRP with n<=1000
6. Release remaining k's to the public for further testing

This was my humble suggestions, but this seems to be the most efficient, however testing large ranges is with current technology bad, when talking about catching the PRP primes turning out to be actually composites. But the listed way, is the most effecient way and less risky of suffering various delaying setbacks. I've suffered many in my first 500M range, but a lot of new scripts has been developed and this really helps making it easier to go with large ranges.

Also a final notice, I've updated my Rb3a website, and Gary it appears that you've either one of your sites (the one with remaining k's) not updated or you have to many primes on your primelist. To crosscheck, I can mention that I've currently 215 primes listed and 973 k's remaining.

Regards

KEP

[KEP]

Originally Posted by henryzz

thanks i think i will reserve a 10M range when i have a core free then
i only have four cores so i tend to not use more than one occasionally two cores per type of work

i have just found another prime
1570340 21918

Welcome on board then... I think if you use sieving after n>1000 or maybe n>2500 that you might be able to do the entire range in 1-2 days in stead of 5 days... just a little trick I've learned during my many hick-ups while running the Riesel base 3 attack for k<=500M... I'm therefor really looking forward, now that I've learned from previous mistakes and setbacks, to see how fast a 1000M range in fact can be done on a Quad.

When you're done with your 10M range, you can just post your remaining k's and I'll add them to my website, aswell remove the k's that can be considered redundant. Hope to see you get just as addicted as I so we can complete level 1 (n<=25K) before the Moon expeditions starts again Anyone up for helping out besides Henryzz?

Regards

KEP

[gd_barnes]

Originally Posted by henryzz

thanks i think i will reserve a 10M range when i have a core free then
i only have four cores so i tend to not use more than one occasionally two cores per type of work

i have just found another prime
1570340 21918

OK, just tell me what k-range you want to reserve. You'd need to post it in the base 3 reservations/statuses thread.


Thanks,
Gary

[gd_barnes]

Originally Posted by KEP

Also a final notice, I've updated my Rb3a website, and Gary it appears that you've either one of your sites (the one with remaining k's) not updated or you have to many primes on your primelist. To crosscheck, I can mention that I've currently 215 primes listed and 973 k's remaining.

Regards

KEP

Hum. I'll have to look up your website and see what the differences are. I balance k's remaining constantly with the mini-drives that are going on so I can't imagine it would be in the k-range that the mini-drives are processing unless either you or me is not totally up to date with the latest primes and k removal.

[gd_barnes]

Originally Posted by gd_barnes

Hum. I'll have to look up your website and see what the differences are. I balance k's remaining constantly with the mini-drives that are going on so I can't imagine it would be in the k-range that the mini-drives are processing unless either you or me is not totally up to date with the latest primes and k removal.

Originally Posted by KEP

Also a final notice, I've updated my Rb3a website, and Gary it appears that you've either one of your sites (the one with remaining k's) not updated or you have to many primes on your primelist. To crosscheck, I can mention that I've currently 215 primes listed and 973 k's remaining.

Regards

KEP

Well...it figures that the thing I don't check has the problem. I always closely check k's remaining as that has the most impact on what is tested in the mini drives and by individuals. The k's that I had remaining were correct at 977 but the # of primes listed should have shown 211 instead of 215. That means my corrected totals of 977+211 = your total of 973+215.

I've now added 7 primes and removed 7 k's per the Riesel base 3 mini drive from the last 12 hours so there are now 970 k's remaining and 218 primes found. But my 970+218 still equals your total of 973+215.

Here is a balancing of my mess :

Prior k's remaining: 977
k's removed from mini-drive today: -7
New k's remaining: 970

Prior # of primes: 215
Miscount of 3 too many primes: -3
One k shown twice with 2 different primes (now removed): -1
Prior primes should have shown: 211
primes added from mini-drive today: +7
New # of primes: 218


My web pages are now up to date with the mini drive as of Nov. 12th at 1:30 AM GMT.

Thanks for checking a part that I hadn't checked yet! :-)


Gary

[KEP]

Originally Posted by gd_barnes

Well...it figures that the thing I don't check has the problem. I always closely check k's remaining as that has the most impact on what is tested in the mini drives and by individuals. The k's that I had remaining were correct at 977 but the # of primes listed should have shown 211 instead of 215. That means my corrected totals of 977+211 = your total of 973+215.

I've now added 7 primes and removed 7 k's per the Riesel base 3 mini drive from the last 12 hours so there are now 970 k's remaining and 218 primes found. But my 970+218 still equals your total of 973+215.

Here is a balancing of my mess :

Prior k's remaining: 977
k's removed from mini-drive today: -7
New k's remaining: 970

Prior # of primes: 215
Miscount of 3 too many primes: -3
One k shown twice with 2 different primes (now removed): -1
Prior primes should have shown: 211
primes added from mini-drive today: +7
New # of primes: 218


My web pages are now up to date with the mini drive as of Nov. 12th at 1:30 AM GMT.

Thanks for checking a part that I hadn't checked yet! :-)


Gary

Glad that I could be of assistance. It now turns out that we both agree on the amount of k's remaining and the amount of k's being primed. Now I just has to update my websites again, so I can get up to pace and balance.

On a side note, ALL k<=419M for Riesel base 3 has now been taken to n<=25000. I expect 1-2 more weeks before completion to k<=500M.

Regards

KEP