Aliquot sequences that start on the integer powers n^i - Page 177

richs · 2022-09-26, 03:47

Releasing 34^105 at index 7. The C139 is ecm'd to t35.

gd_barnes · 2022-09-26, 09:23

A little info. for you Jean-Luc:

I pulled up each sequence on your page starting at the largest base 7420738134810. Of course, the sequences with more iterations take longer because they have to be rebuilt each time. They can take a lot of time but I don't wait for them to come up on my screen. I just keep clicking on the links on your pages. The same-parity ones all come up fast so I quickly hit the back button and go on to the next one.

Going backwards by base, I hit my limit on the 5th base 6469693230 within ~10-15 minutes but it does not give an error/warning message. After hitting my limit, each time I click a link it simply pulls up the page with the count of page requests, IDs created, database queries, CPU (wall clock time) and the Counting since time. It does not display a red warning message like it does when too many IDs are created.

When I started this, all of the above counts were zero. At this moment, it shows 142 page requests, 0 IDs created, 1640372 database queries, and 345.7 CPU (wall clock time). I don't know if I went over the limit on queries or CPU time. I pulled up the counts page a few times throughout the process to get an idea of what might be the limits on some of the categories. My best guess is that I'm limited to 1500000 (1.5M) database queries within an hour. The "Counting since" categories shows the count started at 10:50am site time. At the time of this post, it is ~11:20am site time. So assuming that the limit that I hit is an hourly limit, I'll be able to do this again in ~30 minutes.

It is my thinking that if people keep doing this, you will much more quickly be able to update your pages in a reasonable amount of time.

Another thought: Is there a way for you to set up an automated process to just simply click on the links until it hits a limit, then wait an hour, do it again, etc.? If that could be automated somehow, that would help a lot. I don't think there needs to be any update going on, just access each sequence in order.

gd_barnes · 2022-09-26, 10:13

I think there is a better way. I'm issuing a WGET command to download the ELF files for all sequences in each base. This does not seem to affect the limits but it seems to force the rebuild of each sequence.

I'll let you know how this goes. I've done it with 2 bases so far.

EdH · 2022-09-26, 11:45

Quote:

Originally Posted by gd_barnes

. . .
Another thought: Is there a way for you to set up an automated process to just simply click on the links until it hits a limit, then wait an hour, do it again, etc.? If that could be automated somehow, that would help a lot. I don't think there needs to be any update going on, just access each sequence in order.

That's how I update my local set. I have a script that runs several bases and then sleeps for an hour. I turn it loose and check on it now and then. I'll start it today and see if it helps. Last time it took about 15 hours. I need to add the new bases to my set, anyway. Let's see how long it takes this time and whether I hit a limit. I'm not sure it will help overall, since it only touches the db for local sequences that are still open, but it might help a bit.

gd_barnes · 2022-09-26, 11:58

I have now finished running WGET's on all ELF's for bases > 3000. It works like a charm. All of the sequences are now rebuilt and pull up very quickly. My process is touching all sequences, both open and terminated.

It's easy to tell that the sequences have been rebuilt. Not only do they come up quickly, they no longer have the "Checked, new" under the status for each index.

Jean-Luc, I think you can do a quick update on your pages for all bases > 3000 at this moment.

I was able to run specifically bases 8128 thru 30030 (12 bases/480 sequences) in one big batch. It took 32 minutes to run. I didn't have any hiccups. I'll keep gradually increasing the batch size.

I'll see how far I can get with this before signing off. I'm going to shoot for all bases > 300. Then later on Monday I'll see how many more I can do while gradually increasing the batch size.

EdH · 2022-09-26, 12:47

I've started my local set update as well. You'll probably notice a change as we meet somewhere, but your efforts will be more complete than mine since you are also including terminated sequences.

gd_barnes · 2022-09-26, 13:05

I'm now down under base 800. The large # of double-square bases just above that made it run much more quickly in that area.

I will stop at base 300 for the morning and then do some more starting early evening.

Hopefully after we are done with all of this, Jean-Luc can do all his updates in one setting.

EdH · 2022-09-26, 13:23

My efforts have totally stalled and I'm quite stuck for at least another half-hour. I may have to abandon my update effort for now. It had only made it part way into the base 3 set.

gd_barnes · 2022-09-26, 14:31

I have completed accessing all sequences for bases > 300 using wget. I spot-checked many of the longest sequences to verify that they pulled up quickly. All sequences appear to have been rebuilt.

Jean-Luc, it is my hope that you can quickly run updates for all bases > 300 at this time.

I will continue with this later today.

garambois · 2022-09-26, 16:06

I just got home from work and read your posts.
I think you are right.
I hadn't thought of downloading the .elf's to force the reconstruction of the sequences.
I will do some work in this direction soon so that I can do this myself after the next maintenance operation on FactorDB.
Thank you for helping me in this way.
I will try to make progress with the updates during the week, but it is not easy when I am working.
I will start with the bases >300, it will be a good step done.
For the smaller bases, I'll see later.

EdH · 2022-09-26, 18:48

I tried my update script and it quickly overran the limits twice, without making it through the base 3 table. Apparently just grabbing an .elf is less taxing than checking a last line.

I believe I just verified that as the case. I'll have to rewrite my scripts to stop checking whether work was done and just d/l the current .elf.

2022-09-26, 09:23	#1938
gd_barnes "Gary" May 2007 Overland Park, KS 27414₈ Posts	A little info. for you Jean-Luc: I pulled up each sequence on your page starting at the largest base 7420738134810. Of course, the sequences with more iterations take longer because they have to be rebuilt each time. They can take a lot of time but I don't wait for them to come up on my screen. I just keep clicking on the links on your pages. The same-parity ones all come up fast so I quickly hit the back button and go on to the next one. Going backwards by base, I hit my limit on the 5th base 6469693230 within ~10-15 minutes but it does not give an error/warning message. After hitting my limit, each time I click a link it simply pulls up the page with the count of page requests, IDs created, database queries, CPU (wall clock time) and the Counting since time. It does not display a red warning message like it does when too many IDs are created. When I started this, all of the above counts were zero. At this moment, it shows 142 page requests, 0 IDs created, 1640372 database queries, and 345.7 CPU (wall clock time). I don't know if I went over the limit on queries or CPU time. I pulled up the counts page a few times throughout the process to get an idea of what might be the limits on some of the categories. My best guess is that I'm limited to 1500000 (1.5M) database queries within an hour. The "Counting since" categories shows the count started at 10:50am site time. At the time of this post, it is ~11:20am site time. So assuming that the limit that I hit is an hourly limit, I'll be able to do this again in ~30 minutes. It is my thinking that if people keep doing this, you will much more quickly be able to update your pages in a reasonable amount of time. Another thought: Is there a way for you to set up an automated process to just simply click on the links until it hits a limit, then wait an hour, do it again, etc.? If that could be automated somehow, that would help a lot. I don't think there needs to be any update going on, just access each sequence in order. Last fiddled with by gd_barnes on 2022-09-26 at 09:28

2022-09-26, 11:58	#1941
gd_barnes "Gary" May 2007 Overland Park, KS 27414₈ Posts	I have now finished running WGET's on all ELF's for bases > 3000. It works like a charm. All of the sequences are now rebuilt and pull up very quickly. My process is touching all sequences, both open and terminated. It's easy to tell that the sequences have been rebuilt. Not only do they come up quickly, they no longer have the "Checked, new" under the status for each index. Jean-Luc, I think you can do a quick update on your pages for all bases > 3000 at this moment. I was able to run specifically bases 8128 thru 30030 (12 bases/480 sequences) in one big batch. It took 32 minutes to run. I didn't have any hiccups. I'll keep gradually increasing the batch size. I'll see how far I can get with this before signing off. I'm going to shoot for all bases > 300. Then later on Monday I'll see how many more I can do while gradually increasing the batch size. Last fiddled with by gd_barnes on 2022-09-26 at 12:03

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Broken aliquot sequences	fivemack	FactorDB	46	2021-02-21 10:46
Broken aliquot sequences	schickel	FactorDB	18	2013-06-12 16:09
A new theorem about aliquot sequences	garambois	Aliquot Sequences	34	2012-06-10 21:53
poaching aliquot sequences...	Andi47	FactorDB	21	2011-12-29 21:11
New article on aliquot sequences	schickel	mersennewiki	0	2008-12-30 07:07

2022-09-26, 03:47	#1937
richs "Rich" Aug 2002 Benicia, California 2×823 Posts	Releasing 34^105 at index 7. The C139 is ecm'd to t35.

2022-09-26, 10:13	#1939
gd_barnes "Gary" May 2007 Overland Park, KS 2²×3,011 Posts	I think there is a better way. I'm issuing a WGET command to download the ELF files for all sequences in each base. This does not seem to affect the limits but it seems to force the rebuild of each sequence. I'll let you know how this goes. I've done it with 2 bases so far.

2022-09-26, 12:47	#1942
EdH "Ed Hall" Dec 2009 Adirondack Mtns 2³×677 Posts	I've started my local set update as well. You'll probably notice a change as we meet somewhere, but your efforts will be more complete than mine since you are also including terminated sequences.

2022-09-26, 13:05	#1943
gd_barnes "Gary" May 2007 Overland Park, KS 2²·3,011 Posts	I'm now down under base 800. The large # of double-square bases just above that made it run much more quickly in that area. I will stop at base 300 for the morning and then do some more starting early evening. Hopefully after we are done with all of this, Jean-Luc can do all his updates in one setting.

2022-09-26, 13:23	#1944
EdH "Ed Hall" Dec 2009 Adirondack Mtns 1528₁₆ Posts	My efforts have totally stalled and I'm quite stuck for at least another half-hour. I may have to abandon my update effort for now. It had only made it part way into the base 3 set.

2022-09-26, 14:31	#1945
gd_barnes "Gary" May 2007 Overland Park, KS 2²×3,011 Posts	I have completed accessing all sequences for bases > 300 using wget. I spot-checked many of the longest sequences to verify that they pulled up quickly. All sequences appear to have been rebuilt. Jean-Luc, it is my hope that you can quickly run updates for all bases > 300 at this time. I will continue with this later today.

2022-09-26, 16:06	#1946
garambois "Garambois Jean-Luc" Oct 2011 France 1152₁₀ Posts	I just got home from work and read your posts. I think you are right. I hadn't thought of downloading the .elf's to force the reconstruction of the sequences. I will do some work in this direction soon so that I can do this myself after the next maintenance operation on FactorDB. Thank you for helping me in this way. I will try to make progress with the updates during the week, but it is not easy when I am working. I will start with the bases >300, it will be a good step done. For the smaller bases, I'll see later.

2022-09-26, 18:48	#1947
EdH "Ed Hall" Dec 2009 Adirondack Mtns 2³×677 Posts	I tried my update script and it quickly overran the limits twice, without making it through the base 3 table. Apparently just grabbing an .elf is less taxing than checking a last line. I believe I just verified that as the case. I'll have to rewrite my scripts to stop checking whether work was done and just d/l the current .elf.