mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Aliquot Sequences (https://www.mersenneforum.org/forumdisplay.php?f=90)
-   -   Well that is slightly irritating (https://www.mersenneforum.org/showthread.php?t=23802)

fivemack 2018-11-14 08:08

Well that is slightly irritating
 
The recent factordb upgrade seems to have lost the connectivity information for all the sequences.

So when I try to collect 25 sequences automatically, each of the queries has to recompute the sequence from scratch, and I get thrown out after eighteen for using too much database CPU time. I'm used to getting thrown out for adding too many IDs when submitting my results, but being unable to collect new work in reasonable chunks is somewhat irritating.

yoyo 2018-11-14 21:29

In the background we are going through all open sequences to recreate them in factordb.
But it will need some more days.
On the blue page you can check the last update time of a sequence.

Stargate38 2018-11-15 16:31

You're right about the irritating part. I had the same problem with some sequences starting >10[sup]9[/sup] that I ran, such as 122104081182219999328[sub]10[/sub]=Protoceratops[sub]36[/sub], which has over 6400 terms. It's the longest sequence that I've ever run. It took 3 minutes for it to be recomputed. Something similar happened with the other sequences I've run over the years, such as 4198862272 (even though it has less than 1000 terms).

LaurV 2018-11-16 07:21

Now we know who spams the DB... :razz:

Stargate38 2018-11-17 23:15

It's not me. I only do Aliquot sequences, and they're not that frequent.

LaurV 2018-11-19 05:27

That was a joke, not a real accusation. Therefore the ":razz:" sign.

There is no way for me to know who is spamming the DB, as I am not in the "privileged" list, and I have no more rights than you have, but I am still unhappy about the way the DB is managed, and I won't miss an opportunity to comment about, even if that means upsetting few people/gods here/there. It is not enough to offer the available space and few daemons/elves running factoring jobs and forget about it. Proper maintenance and a user/credential system will solve a lot of these problems. Not any idiot should be allowed to add factors in the db without proper authentication.

On the other hand, queuing aliquot sequences (or any sequence) starting with 20-80 digits have no value, and it will do nothing except loading a lot of small composites in the DB and giving work to the elves (who could otherwise do more important jobs). Such sequences can be done locally with a copy of yafu and aliqueit, where you have the liberty to use pari or perl or a batch file to run a lot of them, they are faster, take only seconds to either terminate or reach a point where they become interesting (like 100 digits whatever) and then, if they are interesting, be reported to the DB.

A sequence with 6400 terms, well... YES, that IS interesting. But other "Protoceratops" sequences are not (haha, buddy, I think you just coined a term!), and I think that such queries should not be accepted by the DB from anonymous users (nothing personal neither talking about you, I was talking generally).

chris2be8 2018-11-20 16:35

At present most of the spam is of the form (14477752*41##+739)/613577 (just look at [URL]http://factordb.com/listtype.php?t=3&mindig=70&perpage=100&start=0[/URL] ). I'm working on numbers in the range from 70-79 digits but can't clear it as fast as it arrives.

If anyone wants to help 78 digit numbers would be a good target.

Chris

Stargate38 2018-11-21 00:54

This is where I got "Protoceratops" from:

[URL]http://www.factordb.com/index.php?showid=1100000001105345291&base=36[/URL]

It's the starting number for that sequence. I usually label my sequences with the base-36 value, but only if it spells out something interesting. Usually I stop a sequence when it hits 120 digits, but there are some exceptions.

richs 2018-11-21 23:17

[QUOTE=chris2be8;500580]At present most of the spam is of the form (14477752*41##+739)/613577 (just look at [URL]http://factordb.com/listtype.php?t=3&mindig=70&perpage=100&start=0[/URL] ). I'm working on numbers in the range from 70-79 digits but can't clear it as fast as it arrives.

If anyone wants to help 78 digit numbers would be a good target.

Chris[/QUOTE]

I pointed a core at the 78 digit numbers early today.

chris2be8 2018-11-22 16:52

Thanks.

Chris

richs 2018-11-25 03:25

[QUOTE=richs;500696]I pointed a core at the 78 digit numbers early today.[/QUOTE]

Approx 2400 78 digit composites factored in 77 hours.


All times are UTC. The time now is 03:27.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.