Certificate : For all indents and purposes - Page 13

Jayder · 2017-02-24, 01:15

I've just taken my last batch of about 500. When I'm done that, there is very little left.

If anybody is still working from 300dd - 3000dd, please post here. But I can handle the last bit without any more help.

EdH · 2017-02-24, 02:45

Quote:

Originally Posted by Jayder

I've just taken my last batch of about 500. When I'm done that, there is very little left.

If anybody is still working from 300dd - 3000dd, please post here. But I can handle the last bit without any more help.

I just uploaded a few that are just above 3000dd, so if your pull took you past 3000dd you might want to cull those ones out. I am only working in the 3000dd and up area now.

I would be interested if you found any more composites, although I suppose one or two might sneak in. I've run my scripts up to 4275dd so far. Interestingly, there have been consistently right at about 20 composites per every 1000 PRPs. Also of interest (to me, at least), is that they are not spread out in any even manner. I'm seeing a pocket often that will have about a 50% ratio.

chris2be8 · 2017-02-24, 17:27

Quote:

Originally Posted by EdH

I'm running an experiment to see if I can handle 5000 PRPs (beginning at 300dd) at once, checking for composites. I'm using YAFU's isprime() test to find them. That step is taking quite some time.

YAFU has a bpsw() test which is probably faster than isprime() and would almost certainly find any composites. According to YAFU's docfile isprime does several rabin-miller tests which might just be duplicating what factordb has already done.

Chris

EdH · 2017-02-24, 18:34

Quote:

Originally Posted by chris2be8

YAFU has a bpsw() test which is probably faster than isprime() and would almost certainly find any composites. According to YAFU's docfile isprime does several rabin-miller tests which might just be duplicating what factordb has already done.

Chris

Thanks Chris,

I was actually just yesterday reading about isprime() being definitive (and therefore more thorough than needed for this application) and wondered if I should consider a different test. I'll give bpsw() a try-out sometime soon.

Ed

henryzz · 2017-02-24, 19:10

I believe that it is an implementation of https://en.wikipedia.org/wiki/Adlema...primality_test

EdH · 2017-02-24, 23:19

Well, I did some testing and bpsw() takes four and one-half times as long to run as isprime(). To verify, I ran tests with the same ten PRPs for both, and isprime() took 50 seconds, while bpsw() took 3 minutes and 56 seconds.

YAFU's isprime() mentions only trial and Rabin-Miller testing. bpsw() mentions the Baillie-Pomerance-Selfridge-Wagstaff probablistic primality test. YAFU has a separate test, aprcl(), that runs the Adleman, Pomerance, Rumely, Cohen, and Lenstra primality test. All this is according to the YAFU docfile.

Jayder · 2017-02-28, 05:00

Just a heads up that I'll be taking everything under 3000dd shortly after my current batch is done, which looks like it will finish in about 20 hours.

If you are working anywhere under 3000dd, speak now.

Jayder · 2017-03-02, 12:20

Finally: for a short time, there were no unproven primes in the database under 3000 digits. The one listed in the screenshot is, in fact, the composite that was discussed a page or two ago that refuses to be recognised as composite by the database.

Good team effort. Maybe we can clear it once a year or something.

EdH · 2017-03-02, 13:51

Quote:

Originally Posted by Jayder

Finally: for a short time, there were no unproven primes in the database under 3000 digits. The one listed in the screenshot is, in fact, the composite that was discussed a page or two ago that refuses to be recognised as composite by the database.

Good team effort. Maybe we can clear it once a year or something.

And, had you looked a little higher, you would have only found one more below 3100.

In addition, the overall number of PRPs is down from >86000 to <81000 ATM.

I think we made a good dent...

pakaran · 2017-03-02, 15:09

I'm continuing to take the smallest numbers, mostly taking a few hundred at a time, one or twice a day. Before essentially everything got finished, I was only rarely getting past 990 dd.

Jayder · 2017-03-02, 21:49

Quote:

Originally Posted by EdH

And, had you looked a little higher, you would have only found one more below 3100.

In addition, the overall number of PRPs is down from >86000 to <81000 ATM.

Nice work!

I am going to move my machine away from this now. I'll do some ECM or something.

2017-03-02, 12:20	#140
Jayder Dec 2012 2×139 Posts	Finally: for a short time, there were no unproven primes in the database under 3000 digits. The one listed in the screenshot is, in fact, the composite that was discussed a page or two ago that refuses to be recognised as composite by the database. Good team effort. Maybe we can clear it once a year or something. Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Fixup Old Primo Certificate?	wblipp	FactorDB	1	2012-05-28 03:16
Invalid certificate?	IvanP	FactorDB	3	2012-05-11 12:17
Could Moore's law be purposely used for marketing purposes?	jasong	Science & Technology	10	2007-01-19 19:04
certificate of appreciation	Unregistered	Information & Answers	13	2004-04-28 06:24

2017-02-24, 01:15	#133
Jayder Dec 2012 2·139 Posts	I've just taken my last batch of about 500. When I'm done that, there is very little left. If anybody is still working from 300dd - 3000dd, please post here. But I can handle the last bit without any more help.

2017-02-24, 19:10	#137
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 2³×3×5×7² Posts	I believe that it is an implementation of https://en.wikipedia.org/wiki/Adlema...primality_test

2017-02-24, 23:19	#138
EdH "Ed Hall" Dec 2009 Adirondack Mtns 111011101001₂ Posts	Well, I did some testing and bpsw() takes four and one-half times as long to run as isprime(). To verify, I ran tests with the same ten PRPs for both, and isprime() took 50 seconds, while bpsw() took 3 minutes and 56 seconds. YAFU's isprime() mentions only trial and Rabin-Miller testing. bpsw() mentions the Baillie-Pomerance-Selfridge-Wagstaff probablistic primality test. YAFU has a separate test, aprcl(), that runs the Adleman, Pomerance, Rumely, Cohen, and Lenstra primality test. All this is according to the YAFU docfile.

2017-02-28, 05:00	#139
Jayder Dec 2012 2×139 Posts	Just a heads up that I'll be taking everything under 3000dd shortly after my current batch is done, which looks like it will finish in about 20 hours. If you are working anywhere under 3000dd, speak now.

2017-03-02, 15:09	#142
pakaran Aug 2002 371₈ Posts	I'm continuing to take the smallest numbers, mostly taking a few hundred at a time, one or twice a day. Before essentially everything got finished, I was only rarely getting past 990 dd.