Sequences with smaller cofactors

Mr. Odd · 2010-11-26, 23:48

I'd like to advance sequences under 10^6 that have cofactors less than 100 digits. Is there an updated list somewhere or an easy way to generate a list from the factor database? Thanks!

schickel · 2010-11-27, 00:38

Quote:

Originally Posted by Mr. Odd

I'd like to advance sequences under 10^6 that have cofactors less than 100 digits. Is there an updated list somewhere or an easy way to generate a list from the factor database? Thanks!

Not by size of cofactor, no, nor is there an overview available from the DB itself.

What size numbers are you willing to tackle? We could probably find you a couple that would be in a position (right now anyway) that they are not going to quickly increase in size. The sequences themsleves will be above 100 digits, but the cofactors would potentially be in the 80-95 digit range.

Mr. Odd · 2010-11-27, 00:48

I can handle cofactors up to 110 digits or so easily though the smaller ones go by faster. I've been looking through this list (http://www.lafn.org/~ax810/five.txt) and checking the currents status of sequences to see if there's any good candidates but that list is badly out of date. If you've got some in mind, by all means post them and I'll get cracking!

kar_bon · 2010-11-27, 00:55

I've written a small script to download all seqs from the FactorDB in ELF-format but it needs about 2 hours for a 100k range, so about 1000 seqs to do so.

So for example the range 100000 to 200000 took 1h 53min for download and the ELF files are about 155 MByte all over.

Here's a list of all open seqs and their last index with incomplete factorization dated from 2010-08-10.
Since then some work has been done in this range and therefore the list is somewhat outdated.

schickel · 2010-11-27, 01:12

Quote:

Originally Posted by Mr. Odd

I can handle cofactors up to 110 digits or so easily though the smaller ones go by faster. I've been looking through this list (http://www.lafn.org/~ax810/five.txt) and checking the currents status of sequences to see if there's any good candidates but that list is badly out of date. If you've got some in mind, by all means post them and I'll get cracking!

OK, here are several from the 200k range that I just downloaded:

Code:

288480  865. sz 101 2^10 * 3^2 * 7^2 * 29
274098  388. sz 100 2^11 * 3 * 7 * 439
272760 1004. sz 103 2^11 * 3^3 * 5^2 * 11

I'm kinda partial to the ones with high exponents for some reason. Typically, the 2 will drop down to some, more reasonable, level within a couple of lines.....

Just run these until they develop a driver and then let us know. I'll mark these as reserved if you like....

bchaffin · 2010-11-29, 22:54

In a similar vein, I realized that there are many sequences <1M which are at less than 100 digits in the DB. For some reason I thought that had all been done, so I've been extending sequences >1M, but I'll work on the smaller ones first if I won't be conflicting with anyone.

What's the right way to go about that? Typically each sequence only takes a few hours, should I just do it silently? Or reserve blocks of 100K or something in the reservations thread? Or something else?

firejuggler · 2010-11-29, 23:13

Your choice
you can use the subproject (taking sequences to 110 digits ) which move at a nice speed, or reserve sequence in the reservation thread.
Or do Both

schickel · 2010-11-30, 07:31

Quote:

Originally Posted by bchaffin

In a similar vein, I realized that there are many sequences <1M which are at less than 100 digits in the DB. For some reason I thought that had all been done, so I've been extending sequences >1M, but I'll work on the smaller ones first if I won't be conflicting with anyone.

What's the right way to go about that? Typically each sequence only takes a few hours, should I just do it silently? Or reserve blocks of 100K or something in the reservations thread? Or something else?

Everything should be up to 100 digits in the DB except sequences that may have been worked by Wieb Bosma. He is/was working the range from 250-450k (and maybe even 450-500k). He has been contacted, but so far has not shared his data.

RobertS was starting to work some of the range covered by Wieb, so he would be the one to contact if you want to co-ordinate. Outside of that, if a sequence is not to 100 digits it's kind of a free-for-all.

If you like, I just pulled the entire 400-500k range a week ago and could send you a list of sequences that were under 100 digits then. As indicated in the other thread, there were also some in the 200-300k that were not up to 100, but those have been worked.... Next up I was going to pull the 300-400k range to see what was going on there.

bchaffin · 2010-12-01, 17:12

OK, thanks schickel. I've done my own DB scan so I already have a list. I'll work on the extending the 500K-1M range to 100 digits for now, and post the blocks I'm working on in the reservations thread. If and when I finish that (which will take a while, especially if I keep pulling all my cores off other tasks to work on 4788!) we'll see whether work remains in the 300K-500K range.

2010-11-26, 23:48	#1
Mr. Odd Mar 2010 5×11 Posts	Sequences with smaller cofactors I'd like to advance sequences under 10^6 that have cofactors less than 100 digits. Is there an updated list somewhere or an easy way to generate a list from the factor database? Thanks!

2010-11-29, 23:13	#7
firejuggler Apr 2010 Over the rainbow 2·1,259 Posts	Your choice you can use the subproject (taking sequences to 110 digits ) which move at a nice speed, or reserve sequence in the reservation thread. Or do Both Last fiddled with by firejuggler on 2010-11-29 at 23:14

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Duplication of Effort for Smaller Aliquot Sequences	EdH	Aliquot Sequences	3	2018-04-17 13:31
Manual Testing ECM on cofactors?	yih117	PrimeNet	24	2018-02-03 15:46
Fermat cofactors	yorix	FermatSearch	10	2017-12-21 16:54
Feasibility of testing Fermat cofactors	JeppeSN	And now for something completely different	6	2017-02-24 10:17
Testing Mersenne cofactors for primality?	CRGreathouse	Computer Science & Computational Number Theory	18	2013-06-08 19:12

2010-11-27, 00:48	#3
Mr. Odd Mar 2010 5×11 Posts	I can handle cofactors up to 110 digits or so easily though the smaller ones go by faster. I've been looking through this list (http://www.lafn.org/~ax810/five.txt) and checking the currents status of sequences to see if there's any good candidates but that list is badly out of date. If you've got some in mind, by all means post them and I'll get cracking!

2010-11-29, 22:54	#6
bchaffin Sep 2010 Portland, OR 7·53 Posts	In a similar vein, I realized that there are many sequences <1M which are at less than 100 digits in the DB. For some reason I thought that had all been done, so I've been extending sequences >1M, but I'll work on the smaller ones first if I won't be conflicting with anyone. What's the right way to go about that? Typically each sequence only takes a few hours, should I just do it silently? Or reserve blocks of 100K or something in the reservations thread? Or something else?

2010-12-01, 17:12	#9
bchaffin Sep 2010 Portland, OR 173₁₆ Posts	OK, thanks schickel. I've done my own DB scan so I already have a list. I'll work on the extending the 500K-1M range to 100 digits for now, and post the blocks I'm working on in the reservations thread. If and when I finish that (which will take a while, especially if I keep pulling all my cores off other tasks to work on 4788!) we'll see whether work remains in the 300K-500K range.