mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Aliquot Sequences (https://www.mersenneforum.org/forumdisplay.php?f=90)
-   -   Aliquot Sequences Summary Page (https://www.mersenneforum.org/showthread.php?t=11874)

kar_bon 2009-05-15 09:14

Aliquot Sequences Summary Page
 
hi folks,

i've just put a page online to make the information about Aliquot Sequences more easier to seek and show:

[url]http://www.rieselprime.de/Others/Aliquot100.htm[/url]

there's not much to see yet:

only 3 of my sequences listed with last edit date, current index (with link to graph) and current line/composite, and the type of driver/guide.

have a look. normally this domain is for my RieselPrimeDatabase k*2^n-1.

Karsten

kar_bon 2009-05-15 10:58

Note:

i've updated that page with some more (about 30 for now) sequences.

will continue.

opinions? suggestions?

btw: i can only obtain data from the FactoringDatabase so some seq's are not on the current status as given in the forum!

10metreh 2009-05-15 11:54

[quote=kar_bon;173645]Note:

i've updated that page with some more (about 30 for now) sequences.

will continue.

opinions? suggestions?

btw: i can only obtain data from the FactoringDatabase so some seq's are not on the current status as given in the forum![/quote]

That is not "about 30".

kar_bon 2009-05-15 12:05

oh, please.....

i've updated this page the last hour and will do more today so don't nail me on this number.

Statistics will be shown later, like number of sequences, digits all over, number of factors, biggest factors, highest exponent of a factor, number of different digits 0 to 9, number of characters show on that page.... :-)

then you have enough numbers to verify for correctness!

THIS page is a first shot so better make some opinions would help me more!

hhh 2009-05-15 12:23

Neat!!!

How about this (my old dream, there is another thread about it): Could you make a retired sequences section, which displays the last unfactored composite, with a reserve button next to it? This would require the users to have some sort of honour thingy not to click if they don't want. Perhaps with a field to fill in ones name. Then display the number as reserved together with the reservation date. I don't know if it is possible as well to program the webpage to contact the database and trigger the ecm very high limits every once in a while for some randomly chosen number (4 times a day?).
But in the meanwhile, just the last composite with link to the database would be great. H.

kar_bon 2009-05-15 21:49

no, this page was done by hand. perhaps Syd can make such a page by php.

just uploaded all seq. of the reservation thread. hope there're no errors.
not yet included the ones from the retired, recommended and first holes 230-232k thread.

10metreh 2009-05-17 10:19

You seem to have called 2^3*5 a driver. It is not. 2^3*3 and 2^3*3*5 are.

kar_bon 2009-05-17 11:58

corrected.

i've also deleted the link for the 'Current Composite' because it's available under the 'Size' column, so easier to update for me, too!

Andi47 2009-05-17 12:53

nice!

The Lehmer Five are marked as unreserved - aren't they still reserved by P. Zimmermann?

10metreh 2009-05-17 13:19

[quote=Andi47;173894]nice!

The Lehmer Five are marked as unreserved - aren't they still reserved by P. Zimmermann?[/quote]

They seem to be reserved by P. Zimmermann on the new page to me - is there something wrong with your browser?

Andi47 2009-05-17 13:59

[QUOTE=10metreh;173897]They seem to be reserved by P. Zimmermann on the new page to me - is there something wrong with your browser?[/QUOTE]

The status of "reserved", "available" and "terminated" is color-coded: yellow for reserved and light blue for available.

The Lehmer Five are saying "last editor: P. Zimmermann", but they are color-coded in light blue (meaning "available"). I checked with firefox and IE.

10metreh 2009-05-17 15:29

[quote=Andi47;173902]The status of "reserved", "available" and "terminated" is color-coded: yellow for reserved and light blue for available.

The Lehmer Five are saying "last editor: P. Zimmermann", but they are color-coded in light blue (meaning "available"). I checked with firefox and IE.[/quote]

Ah, I see now. :smile:

kar_bon 2009-05-17 19:53

1 Attachment(s)
i updated the Lehmer 5 reserved by P.Zimmermann (and 4788 and 130396).
the data i got was out of date and i was not sure if he works on it further.

need some time to make all things work/look fine because i'm not so familiar with drivers/guides or the current reservations (only out of this forum so far).

please post here, if there're any other errors.

attached a view of that page it should look like this. if not there's something not correct with the browser settings.

schickel 2009-05-18 03:01

[QUOTE=kar_bon;173926]i updated the Lehmer 5 reserved by P.Zimmermann (and 4788 and 130396).
the data i got was out of date and i was not sure if he works on it further.[/QUOTE]I think they're currently low priority. If there were the interest, the forum could adopt them, too.[quote]need some time to make all things work/look fine because i'm not so familiar with drivers/guides or the current reservations (only out of this forum so far).[/quote]Did you check the wiki article? It covers the main points about the drivers; just ask here if you have any questions.....

As far as reservations, to the best of my knowledge, we're alone in the field in the 100-200k range. Check this [url="http://www.mersenneforum.org/showthread.php?t=11625"]thread[/url] for the status of other ranges.....[quote]please post here, if there're any other errors.[/quote]I didn't see anything right off hand. Nice work on the table.....[quote]attached a view of that page it should look like this. if not there's something not correct with the browser settings.[/QUOTE]

10metreh 2009-05-18 17:12

Errrm... 2^2*7 is not the downdriver.

schickel 2009-05-20 04:48

Couple of things
 
[b]4788 - 2[sup]7[/sup] * 3[sup]2[/sup][/b] is a guide, not a driver.

[b]29772 - 2[SUP]4[/SUP] * 7[/b] is a guide.

[b]103296 - 2[SUP]2[/SUP] * 3[/b] is a guide.

[b]112638 - 2[SUP]2[/SUP] * 3[SUP]2[/SUP][/b] is a guide

[b]115302 - 2[SUP]6[/SUP] * 3[/b] is actually controlled by [b]2[SUP]6[/SUP] * 127[/b]

[b]139314 - 2[SUP]5[/SUP] * 7[/b] is a guide.

[b]150096 - 2[SUP]4[/SUP] * 3[/b] is a guide.

[b]162126 - 2[SUP]2[/SUP] * 3[SUP]2[/SUP] * 5[/b] is a guide.

axn 2009-05-20 05:32

[QUOTE=schickel;174242][b]4788 - 2[sup]7[/sup] * 3[sup]2[/sup][/b] is a guide, not a driver.
<snip>[/QUOTE]

Hmmm... Either my definition of guide is messed up or most of these are incorrect. What's the definition of driver/guide you're using?

[code]
4788 2^7*3^2*17 Guide
9708 2^2*7 Driver
10212 2*3 Driver
29772 2^4 Guide
88662 2^2 Guide
100320 2^2*7 Guide
100436 2^4*31 Guide
103296 2^2 Guide
112638 2^2 Guide
115302 2^6*127 Driver
123960 2^2 Guide
128370 2^2 Guide
129336 2^2 Guide
132792 2^4*31 Driver
132920 2^4*31 Driver
139314 2^5*7 Guide
145008 2*3 Driver
149808 2^2*7 Driver
150096 2^4 Guide
162126 2^2 Guide
167148 2^4 Guide
168912 2^4 Guide
171018 2^2 Guide
175410 2*3^2 Driver
182224 2^4*31 Driver
[/code]

schickel 2009-05-20 05:52

[QUOTE=axn;174247]Hmmm... Either my definition of guide is messed up or most of these are incorrect. What's the definition of driver/guide you're using?
[/QUOTE]Drivers are defined in this article:[quote]R. K. GUY and J. L. SELFRIDGE, "What drives an aliquot sequence," Math. of Comp. v. 29, n. 129, 1975, pp. 101-107.[/quote]I'm not sure if my use of guide is correct, but I use "guide" for anything that persists, but that's not a driver.

The drivers outlined in the above article are:[code][tex]2[/tex] (the downdriver)

[tex]2^2 * 7[/tex]

[TEX]2^3 * 3[/TEX]

[TEX]2^3 * 3 * 5[/TEX]

[TEX]2^4 * 31[/TEX]

[TEX]2^5 * 3 * 7[/TEX]

[TEX]2^6 * 127[/TEX]

[TEX]2^9 * 3 * 11 * 31[/TEX][/code]

10metreh 2009-05-20 06:25

[quote=schickel;174250]but I use "guide" for anything that persists, but that's not a driver.[/quote]

What is the word Guy and Selfridge suggest using for things like 2^2*3, which would be classified as having the "guide" 2^2 (seems completely mad)?

schickel 2009-05-20 06:47

[QUOTE=10metreh;174256]What is the word Guy and Selfridge suggest using for things like 2^2*3, which would be classified as having the "guide" 2^2 (seems completely mad)?[/QUOTE]Here's the definition from the article:[quote=Guy&Selfridge]Define a [U]guide[/U] to be [tex]2^a[/tex], together with a subset of the prime factors of [TEX]\ \sigma(2^a)[/TEX]. A [U]driver[/U] is defined as a number [TEX]\ 2^av[/TEX] with [TEX]\ a>0[/TEX], [TEX]v[/TEX] odd, [TEX]v|\sigma(2^a)[/TEX] and [TEX]2^{a-1}|\sigma(v)[/TEX]. The last requirement is included so that the power of the prime 2 will tend to persist at least as well as it does for the driver 2 itself, for which the condition is trivially satisfied.[/quote]Further down, they give these as guides:[code][TEX]2^2[/TEX]

[TEX]2^3[/TEX]

[TEX]2^35[/TEX]

[TEX]2^4[/TEX]

[TEX]2^53[/TEX]

[TEX]2^57[/TEX]
[/code]So, I have been a little, um, imprecise, in calling, for example, [TEX]2^23[/TEX], a guide. The guide is actually [TEX]2^2[/TEX]...

And I actually have been lax: the drivers include the [i]even perfect numbers[/i].....

axn 2009-05-20 06:57

[QUOTE=schickel;174250]Drivers are defined in this article:I'm not sure if my use of guide is correct, but I use "guide" for anything that persists, but that's not a driver.
[/QUOTE]

Ok. I get one definition of guide from [URL="http://www.emis.de/journals/EM/expmath/volumes/11/11.2/3630finishes1.pdf"]here[/URL] (PDF file).

Basically, what forms a driver or guide depends on the factorisation of Sigma(2^a).

EDIT:- Frank was here first :)

schickel 2009-05-20 07:07

[QUOTE=axn;174265]Ok. I get one definition of guide from [URL="http://www.emis.de/journals/EM/expmath/volumes/11/11.2/3630finishes1.pdf"]here[/URL] (PDF file).

Basically, what forms a driver or guide depends on the factorisation of Sigma(2^a).

EDIT:- Frank was here first :)[/QUOTE]Yes, in the 3630 article, they give a quick review of the definition. If you're interested in the full proof, you have to go back to the Guy & Selfridge article.

If you have access to a library with inter-library loan privileges with a school, you can request a copy of the original article. (I got mine from a co-worker who was moonlighting from their day job at the local community college....)

kar_bon 2009-05-20 08:14

corrected.

is it a 'must have' of such colors/definitions? tell me your opinion.

what about this:

i cancel the different coloration for drivers/guides, but only put the 'small' factors in. if Syd can build an automated page like this, i think it would be a tricky algorithm to decide if there's a drive or guide in the sequence!

another aspect:
i got not the time to go trough all sequence to update them every day, but i can do this when changes posted in this forum.

schickel 2009-05-20 08:43

[QUOTE=kar_bon;174276]corrected.

is it a 'must have' of such colors/definitions? tell me your opinion.[/QUOTE]No, not really. It gets to be where you can tell at a glance whether it's a driver or not.[quote]what about this:

i cancel the different coloration for drivers/guides, but only put the 'small' factors in. if Syd can build an automated page like this, i think it would be a tricky algorithm to decide if there's a drive or guide in the sequence![/quote]That's pretty much what I do, just put the small factors in.[quote]another aspect:
i got not the time to go trough all sequence to update them every day, but i can do this when changes posted in this forum.[/QUOTE]There's no sense in doing it everyday, especially with things very possibly changing every day. I was updating the reservations every couple of days at the start, but since we've all settled down in to the long grind part of things, weekly is it right now, unless someone posts about a great big ole downdriver run.

kar_bon 2009-05-20 09:31

updated the page with:

- no colors for drivers/guides (easier for me too)

- inserted the "Godwin twelve": sequences between 1000 and 2000
i reserved them by C.Clavier although on his page are some others mentioned

more to come and then splitting on more pages (perhaps all from 0-100000, 100000-200000,...).
i noticed there're about 1000 sequences per 100000.

kar_bon 2009-05-22 06:42

new updates are online:

- in post #1 the link was changed (thanks Frank) because of splitting the ranges for sequences:
now the start-link ist the range 100000 <= n < 200000, the main range for this project.

- a link for 2 <= n < 100000 is available on this page to (and vice versa on the other page)

i tried to fill in some 'old' results (ended sequences) in the lower ranges.

more to come.

kar_bon 2009-05-26 21:14

news:

- made a new page for 200000 <= n < 300000 sequences (from first holes thread here)

- included all n<10000 (mostly with only short link on bottom of the page)

kar_bon 2009-06-02 11:06

just updated the 100000 < n 200000 page.

notes:

- please Frank, update the FactorDatabase for 163716 (last terms not in yet, note on page)

- 144984 now at index 7039

- highest sequence from Markus (Syd): 162126 at 153 digits!


question:
does anybody uses the link to the latest unfactored term of a sequence (on the page the column 'Size')?

if not, i can remove those links and the update will be quicker!

(the link for the sequence itself popup the main screen of it in the database, so all other links are redundant, but i think the separate link to the graph is quite useful)

Karsten

axn 2009-06-02 11:12

[QUOTE=kar_bon;175614]does anybody uses the link to the latest unfactored term of a sequence (on the page the column 'Size')?[/QUOTE]

Probably not. I mean, it adds very little value.

BTW, do you do the updates manually? Can you use some kind of HTML parsing and fetch/update the results automatically by loading the relevant pages from Syd's database?

Probably, Syd can create a HTTP/XML interface that just gives the vital statistics of a given sequence.

kar_bon 2009-06-02 11:16

[QUOTE=axn;175616]BTW, do you do the updates manually? Can you use some kind of HTML parsing and fetch/update the results automatically by loading the relevant pages from Syd's database?[/QUOTE]

i do this manually: click on the sequence and update the html with the new values.

[QUOTE]
Probably, Syd can create a HTTP/XML interface that just gives the vital statistics of a given sequence.[/QUOTE]

i hope so too!
a page created by a php script would be easiest, so 'calling' such a page would have the latest results (if they're in database!).

Greebley 2009-06-05 17:57

It appears that 4350 is missing from the summary list.

Edit: And 4380.

They are on Clavier's page

Edit: 4800, 4842. The list isn't supposed to be complete?

kar_bon 2009-06-05 19:13

oh, look at the end of the table.
i inserted some seqs from bottom (only numbers with link) but not sorted yet!

the n<100000 is not complete yet (only n<10000 all shown). just now i'm working on the [b]complete[/b] lists for 100000<n<200000 and 200000<n<300000!

kar_bon 2009-06-05 22:56

bigger update:

- 0 < n < 100000: all seqs < 10000 are included with current data from Factoring Database

- 100000 < n < 200000: [b]all[/b] seqs included (thanks Frank for the data), not yet validated with FDb, reserved seqs with latest data from FDb

- 200000 < n < 300000: [b]all[/b] seqs included (thanks Frank for the data), not yet validated with FDb, most seqs > 220000 are without sizes and no data in FDb!

10metreh 2009-06-06 11:10

[quote=kar_bon;176146]- 0 < n < 100000: all seqs < 10000 are included with current data from Factoring Database[/quote]

You can get the status of 10000-100000 [URL="http://www.lafn.org/~ax810/five.txt"]here[/URL]. Note that [code]12012 1710. 2 C97 sz 97[/code] is wrong: this was the old status of 10212, and 12012 was a typo that was never removed. Note also that 11280 has its old status from before when a typo was discovered in the earlier work.

Andi47 2009-06-06 11:48

[QUOTE=10metreh;176188]You can get the status of 10000-100000 [URL="http://www.lafn.org/~ax810/five.txt"]here[/URL]. Note that [code]12012 1710. 2 C97 sz 97[/code] is wrong: this was the old status of 10212, and 12012 was a typo that was never removed. Note also that 11280 has its old status from before when a typo was discovered in the earlier work.[/QUOTE]

Speaking of 10212 - I have picked up this one again and will do a few more iterations, but I don't plan to take it very far... (well... I might change my mind if it gets rid of that 2*3 driver.)

Greebley 2009-06-06 14:46

Thanks for updating the list!

I did notice for some of the 10000 - 100000 range were not in the database. For example, [URL="http://factorization.ath.cx/search.php?query=&se=1&aq=10344&action=last&fr=&to="]10344[/URL] only goes to 74 digits. Is this because someone else is doing the 10000 -100000 range? I was curious about what was the lowest sequence for which work done is NOT in the database, and 10344 appears to be the answer.

10metreh 2009-06-06 15:14

[quote=Greebley;176214]For example, [URL="http://factorization.ath.cx/search.php?query=&se=1&aq=10344&action=last&fr=&to="]10344[/URL] only goes to 74 digits. Is this because someone else is doing the 10000 -100000 range? I was curious about what was the lowest sequence for which work done is NOT in the database, and 10344 appears to be the answer.[/quote]

It actually goes to 78 - the size of the initial composite is not the size of the remaining composite.

Clifford Stern maintains the 10000-100000 range, and he is constantly pushing the sequences in it higher. However, he has not put all that work into the database.

kar_bon 2009-06-08 15:29

now all seqs for 10000 < n < 100000 are shown on the Summary pages, too.

but: the listed depths are mostly not in the database, yet.

same for n>250000.

so, i think, for a first milestone, all open (and some ended) seqs for n < 300000 are shown on those 3 pages!

next steps:
- check data with database
- insert more seqs in database
- stats/odds page
- try to 'hire' Clavier, Stern, Zimmermann, Bosma, Creyaufmueller and other contributors for this forum/database to share their results

suggestions?

Greebley 2009-06-08 20:44

1 Attachment(s)
Not sure if this is useful, but it is a list of the numbers that reach 1 quadrillion(10^15) before the merge with an earlier sequence or terminate.

The list is of the form:
<number> | <maximum> | <merged or maxed>

The maximum is pretty low. I used 8 byte numbers with trial division up to 2^26 from a program I wrote before finding this group. However, since 1 quadrillion is less than the square of 2^26, all sequence reached 1 quadrillion. In fact they only stopped when going over 2^64, or some factor was greater than the max prime squared and no prime divided it.

I included this because it has the 3740 numbers that might be significant and includes ones that may have been weeded out because they merged at some point after 1 quadrillion. I don't know if you want to put these numbers into the list, or whether you already have this info, but I figured it wouldn't hurt to mention it.

Out of the 3740, 113 merged with a previous sequences given the limits above.

I also have a list of:

Those that reached 1 trillion, but not 1 quadrillion. (340 numbers)
Those that reached 1 billion, but not 1 trillion. (662 numbers)

Greebley 2009-06-08 20:45

1 Attachment(s)
Here is the numbers that reach 1 trillion but not 1 quadrillion.

Greebley 2009-06-08 20:46

1 Attachment(s)
and the one that reach 1 billion but not 1 trillion.

kar_bon 2009-07-04 07:32

Help needed!
 
i'm trying to complete the page for 100k<n<200k sequences.
but there're a lot of them not yet filled with the whole seq in the database from Syd.

so, please, anyone can help here!

to do:
- go to the summary page [url=http://www.rieselprime.de/Others/Aliquot100.htm]here[/url]
- (by now all seqs upto 122k are filled in) click on a seq-number
- compare the given parameters from the summary page with the database

- if the database-index is smaller, do:
- download a seq-file from Franks page [url=http://www.frontiernet.net/~aliquot/alq/100k/]here[/url]
- convert the *.ALQ file into a textbased *.ELF file with [url=http://www.aliquot.de/archiv/alq.zip]ALQ2ELF.EXE[/url]
- upload from missing index to the database
- sometimes the database needs a TF to 10^7 to bring the composite cofactor like in the summary noted!

i will try to fill in more, but it's very timeconsuming (got all those elf) to upload them.

if a file is in the database, i need only seconds to verify against the summary page.

after all seqs are in the database and the summary is checked, i could include some stat there, too:

max/min/mean number of indices/sizes/composites so everyone can see the progress in this range!

10metreh 2009-07-04 11:06

Karsten, 157752 is reserved by D. Stevens (aka henryzz). You seem to have missed it out.

Mini-Geek 2009-07-04 12:20

Let's do some reservations to help make the work filling in the DB a bit more orderly:
Reserving 122540-130k (122540 is the first after Karsten's indicator on the page)
Now if anybody wants to help work on it while I'm still not up to 130k, you can just start at 130k. :smile:
Edit: Do we need to do anything when the DB has a newer status than your page? e.g. [URL="http://factorization.ath.cx/search.php?aq=123450"]123450[/URL]

kar_bon 2009-07-04 12:50

[QUOTE=10metreh;179749]Karsten, 157752 is reserved by D. Stevens (aka henryzz). You seem to have missed it out.[/QUOTE]

included!

[QUOTE=Mini-Geek;179754]Let's do some reservations to help make the work filling in the DB a bit more orderly:
Reserving 122540-130k (122540 is the first after Karsten's indicator on the page)
Now if anybody wants to help work on it while I'm still not up to 130k, you can just start at 130k. :smile:
Edit: Do we need to do anything when the DB has a newer status than your page? e.g. [URL="http://factorization.ath.cx/search.php?aq=123450"]123450[/URL][/QUOTE]

cool, thanks!

no, nothing you can do.
when i'm processing through the summary i'll update it there.

so i'll continue from n=130k!

PS: if anybody finds a higher range in the database other than in the summary (only for the checked ones, not reserved ones, so n<122k) please post here!

Mini-Geek 2009-07-04 13:32

I finished the work for my reservation, now I'm waiting for the DB to finish processing it all so I can double check that it's all in there and processed without any squared line bugs or anything. :smile:
That'll take a while...(the processing, I mean)
I'll take 130k-140k.

Here's a good way I found to do it, it's pretty efficient: (for you or anyone else that might not be doing it very quickly)

Keep these open:
Karsten's summary, the alq index, the Report Factors page, and a command line window at alq2elf's position.

Loop through until you finish: {
Loop through until you've got 5 windows open at the end of the loop: {
click the next item in Karsten's summary and compare the index, and if the index matches then the composite size
if the index is already correct, make sure the composite is correct, and close the window
if it's not, leave the window open
}
download the 5 alq's to your alq2elf location
run alq2elf on each
use the batch file upload at the Report Factors page to begin uploading them all
}

kar_bon 2009-07-04 13:38

ok, continue with n>130k then.

i'm just checking/updating at n=125k!

that's what i'm doing:
- open the summary page as html and text file
- click a seq
- update the textfile

after all is done, upload the new html!

oh, yeaaaahh, i know,.... by php it would be a 'little' bit faster :grin:
but who will write that?

Mini-Geek 2009-07-04 14:21

130k-140220 done, (accidentally went 2 farther than 140k) waiting for processing. I made this batch file:
[code]wget http://www.frontiernet.net/~aliquot/alq/100k/%1.alq
alq2elf %1
wget http://www.frontiernet.net/~aliquot/alq/100k/%2.alq
alq2elf %2
wget http://www.frontiernet.net/~aliquot/alq/100k/%3.alq
alq2elf %3
wget http://www.frontiernet.net/~aliquot/alq/100k/%4.alq
alq2elf %4
wget http://www.frontiernet.net/~aliquot/alq/100k/%5.alq
alq2elf %5[/code]To use it, you can save it as e.g. dl.bat, then run e.g. "dl 139650 139830 139854 140184 140220" and it will download and convert (but not upload) those 5. Note that this must take exactly 5 inputs, and alq2elf and wget must be available from that folder. Made in Windows XP, should work in Linux, too, with little to no tweaking. It's only been used/tested once, (I made it near the end of my range) but seemed to work just fine. :smile: Plus it's dead simple...there's not too much to go wrong besides you typoing a number. Anyone know if a command using wget could be used to upload the file as well? I wouldn't want to use aliqueit's upload function because uploading this much in the foreground would take forever!

I'll take 140k-150k, too. I wonder how long this backlog is going to take to clear out...

kar_bon 2009-07-04 14:31

found a 'missing' seq: 126600 is on Frank's page as ALQ-file but not listed in the status-file i used to compile the summary-page.

i uploaded this file, too, and inserted in the summary page (upload later today).

Mini-Geek 2009-07-04 16:02

140k-150k complete (only took me 25 minutes for the 93 sequences! that's an average of about 16 seconds/sequence :showoff:)
150k-160k reserved

kar_bon 2009-07-04 16:08

currently seq 128808 is primality-proofing by the workers, so don't overload the database 'entrance'! :grin:

Mini-Geek 2009-07-04 16:39

150k-160k complete.
160k-180k reserved.

kar_bon 2009-07-04 16:45

so you will finish the next 'minutes' to n=200k!!!

i've uploaded the current summary pages with some blue marked seqs, to remind me they have to be uploaded when i got the chance to do so!

kar_bon 2009-07-04 17:54

tried to update further, but this issue occurs:

while primality-proofs for seqs in the n=140k range is done, many seqs smaller than these are not yet in the database!

i've uploaded the current 100k-page.

please, Timothy, check this: all seqs for 129k < n < 132k without todays date are not in the database yet!
perhaps i have to wait much longer? or these were not correctly submitted or inserted in the database, because of so much amount of income!?

Mini-Geek 2009-07-04 18:29

[quote=kar_bon;179786] please, Timothy, check this: all seqs for 129k < n < 132k without todays date are not in the database yet!
perhaps i have to wait much longer? or these were not correctly submitted or inserted in the database, because of so much amount of income!?[/quote]
Hm...I'm not sure. If they don't get processed after the rest finish, I'll resubmit them. I still have all the .elf files that were downloaded and converted. Those files appear to be in good condition. I just resubmitted the first 5 after your break point (129k), and then resubmitted the first one (129168), still using the file upload form, and 129168 got processed right after the current seq finished processing, the rest haven't been processed yet. I don't know what's up.

It'd be nice to see a queue for queued up files/factors with, if known, what sequence or special form it's from. A Run Now button would be good too. Also, it'd be good if other workers, or at least multiple cores on the server, could be assigned to process ranges and verify the factors, or whatever the server does.

By the way, I'm done through 170k, paused a while ago, resuming soon.

127980 has the first 6 columns in green, then the last in teal. Also it's only at index 406 in the DB.

kar_bon 2009-07-04 18:36

[QUOTE=Mini-Geek;179789]By the way, I'm done through 170k, paused a while ago, resuming soon.

127980 has the first 6 columns in green, then the last in teal. Also it's only at index 406 in the DB.[/QUOTE]

yes, i saw 129168 prime-proofing, too and it's fully in database now!

ok then, go on.

i remove this color, when all is done (only a reminder for now)!

Mini-Geek 2009-07-04 19:53

170k-187k finished
A majority of the ones over 175k (I'd guesstimate 70%, I'll get more accurate figures after I finish to 200k :smile:) are already filled in Syd's DB. I don't know why. Continuing, I'll edit this post when I finish...
Edit: 187k-200k finished! Looks like the processing is still back at 172032ish. For 175k-200k I uploaded 71 seqs out of 225, or 31.555...%, meaning 68.444...% was the exact figure. Looks like my guess was pretty darn close!

gd_barnes 2009-07-05 09:14

Karsten, you didn't have any size or factor size on 147960. I saw why. There was nothing in the DB for it. Therefore I set the DB to run it as far as it could and it got up to i=487. I then ran aliqueit.exe and filled it in for i=488-662, which is what you had it at. At that point, it encountered a hard C101 so I am not reserving it. Official status:

147960, i=663, sz 104, C101, finished

BTW, I've been trying to upload factors for indexes on several different sequences today. No luck. It does nothing. Also, in the last couple of hours, I've trying doing "ECM very high limits" on multiple different sequences. No luck. It gets through the "preparing for ECM very high limits" phase and then just kicks it right back out without starting. It was also acting a little flaky yesterday towards the end of the day. I also noticed that there are 14 jobs in the queue for ECM very high limits. That is abnormally low. It's almost always been between 30 and 50 ever since I started on the sequences so clearly there are many people that are experiencing the same problem.

Something is clearly wrong with the database right now.


Gary

kar_bon 2009-07-05 09:38

yesterday, Mini-Geek was helping me to update and upload [b]all[/b] seqs not yet in the database for 100k<n<200k.
go to the workerstatus-page and pick a factor which is currently in primality proof.

just now i got one from seq 141930!

so the DB is still inserting those high amount of seqs uploaded from yesterday!

so still wait 'some' time for the seqs, don't know how long!

i've noticed the low amount of VHL-ECM, too, but there's an i7 online since yesterday.
perhaps the queue become lower therefor!?

Andi47 2009-07-05 09:45

[QUOTE=kar_bon;179817]yesterday, Mini-Geek was helping me to update and upload [b]all[/b] seqs not yet in the database for 100k<n<200k.
go to the workerstatus-page and pick a factor which is currently in primality proof.

just now i got one from seq 141930!

so the DB is still inserting those high amount of seqs uploaded from yesterday!

so still wait 'some' time for the seqs, don't know how long!

i've noticed the low amount of VHL-ECM, too, but there's an i7 online since yesterday.
perhaps the queue become lower therefor!?[/QUOTE]

Perhaps there is some kind of bug? I clicked "ECM high limits" for a c128 cofactor and the db reported to have done 120 curves at B1=90k after only 5 seconds! I can't believe this short time...

Edit: Hmmmm strange now, two minutes later, it reports to have split off a p14 factor.

kar_bon 2009-07-05 09:51

[QUOTE=Andi47;179818]Edit: Hmmmm strange now, two minutes later, it reports to have split off a p14 factor.[/QUOTE]

what i meant! perhaps an OC i7! don't know which guy.

10metreh 2009-07-05 10:52

[quote=gd_barnes;179816]Karsten, you didn't have any size or factor size on 147960. I saw why. There was nothing in the DB for it. Therefore I set the DB to run it as far as it could and it got up to i=487. I then ran aliqueit.exe and filled it in for i=488-662, which is what you had it at. At that point, it encountered a hard C101 so I am not reserving it. Official status:

147960, i=663, sz 104, C101, finished
[/quote]

The info on Karsten's page comes from [URL="http://www.frontiernet.net/~aliquot/100000.txt"]this[/URL] list on Frank Schickel's page, not the database. It shows just [code]147960 663. 2^2 * 7[/code] with no size or composite size for 147960. There are a few with no composite size that still have the sequence size. I have a feeling there are a couple with a composite size but no sequence size. Frank last updated this list in May, so it is out of date.

kar_bon 2009-07-05 11:01

i hope, i can update the summary page totally today.

'only' 425 seqs to go. DB is still PrimeProofing at n=144k seqs.

kar_bon 2009-07-05 12:20

320 seqs to go for checking/inserting in DB!

gd_barnes 2009-07-05 22:00

Trust me guys, the database is having problems. Lots of problems. There are plenty of workers. I've seen it run just fine with less workers than it has now.

I do a lot of work with it. I have a job just stuck on 'ECM medium limits', a problem that I've experienced before. Multiple jobs just keeping kicking out after trying 'ECM very high limits' and going through their 'Preparing for ECM very high limits' phase. Sure, they seem to eventually run but instead of kicking out, which causes me to run them several times before giving up, they should show 'in queue for very high limits'. Try it for yourself and see. On occasion, they will run right away like they should but it's quite rarely right now.

I can understand the queue-up on the huge # of factors that you guys sent and my one sequence did finally have its factors inserted but the problems are up and above there not being enough workers. It's not showing 'ECM very high limits' jobs in the queue, even though it appears that some of them may be just that, but in the background.

Syd has been mostly unresponsive in another thread about other problems so I'm not sure what to do at this point. I checked a couple of days ago and he hadn't been on the forum in over a week.

Later today, I'll have a status update. I've completed a fair amount of work since the last status including catching my 1st decent downdriver run on one of my sequences. I lopped over 20 digits off before losing it followed by losing another 10 digits after gaining it back and then losing it again but it was a fun ride! I've now added over 440 indexes to that one starting from a size of 100. :smile: Considering that all I'm really doing when extending these things is looking for sequences that have a low factor or size with their last status and not really looking at the current drivers or guides on any of them, I'm happy when I can catch even 1 decent downdriver out of 30-40 of them.


Gary

gd_barnes 2009-07-05 22:05

[quote=10metreh;179822]The info on Karsten's page comes from [URL="http://www.frontiernet.net/~aliquot/100000.txt"]this[/URL] list on Frank Schickel's page, not the database. It shows just [code]147960 663. 2^2 * 7[/code] with no size or composite size for 147960. There are a few with no composite size that still have the sequence size. I have a feeling there are a couple with a composite size but no sequence size. Frank last updated this list in May, so it is out of date.[/quote]


Yeah, I was aware of that. That is why I went ahead and filled it in. We had no actual factoring data to back up the status.

IMHO, reporting a status without either providing a complete factoring file (.ELF or substantially similiar file) or entering the same in the database is not really a status. Somebody could just make up where they are at or have an error in the factoring sequence somewhere. They could even be off by 1 or 2 indexes or have missed one of the driving factors or have an extra one. That's not a hit on anyone's integretity here or elsewhere. It's just that it's very possible to have errors creep in without supporting information for a status.

That's my 2 cents anyway. :smile:


Gary

kar_bon 2009-07-05 22:11

so about 160 seqs to fill in the database and the range 100k<n<n200k is complete with all factors available from Frank's page.

[b]that's[/b] the status we can work from and the summary page for that range will contain all those results, too.

hope, there're no (not much) typos in; if any, inform me please!

PS: before i forget it: thanks Mini-Geek for the upload, it saved me much time. so this time it was a Maxi-Geek! :grin:

kar_bon 2009-07-06 00:03

[b][color=red]status:[/color][/b]

summary page for 100k<n<200k is checked upto n=170k with database and online.

remaining seqs have to wait until tomorrow! :yawn: :sleep:

gd_barnes 2009-07-06 02:05

Karsten, you are the true die hard. :smile::flex::bow: Thanks for getting all of that filled in. Also thanks to Tim (Mini-Geek) for the assistance.

On another note: The database is working great now. But I'm 80-90% sure that I know exactly what was happening. It was as a result of the huge amount of factors being uploaded all at once but the symptoms were so confusing that I was unable to correlate the two for a long time. But just a few hours before the uploading finished, I was finally able to isolate the symptoms and draw a strong conclusion because the evidence was quite obvious when observed in total. I figured I would give everyone my analysis as to what happened in case it happens again. Here were the symptoms that happened as I tried to extend various indexes:

1. I would trial factor to 1e7. After a normal amount of time, the screen would return with the correct factors...no problem there.
2. I would do ECM low limits 2 times. I always run low-medium limits twice before progressing. After a normal amount of time, the screen would return with ZERO factors on both tries. This happened every time! (It took me a while to figure out that that was happening.)
3. I would do ECM medium limits 2 times. Like #2, the screen would return with zero factors on both tries. Once again, it happened every time.
4. I would do ECM high limits 3 times. Like #2 and #3, the screen would return with zero factors all 3 times every time.
5. I would do ECM very high limits. After it went through its "Preparing for ECM very high limits" phase, it would just stop BUT...not all of the time. A small percentage of the time, it would continue.

As maddening as this was, I became more concerned than frustrated because I thought that the DB was missing factors. I also thought it might be something with the refreshing of the screen while I was connected to the site. So I exited out, even shut off my machine in case something was filling up, turned it back on, reconnected, and still no luck...no small factors.

Then I had a hunch: I went back to the site after about 30-45 minutes and checked them again. Alas, there they were...the small factors that the database "appeared" to miss. About the problem with it not continuing with "ECM very high limits": After analyzing several occurrences of this, I was able to determine that the only time that it would not keeping running that phase is because it had already found smaller factors "internally" but for some reason would not show them on the screen.

Conclusion:
If you see a whole bunch of primality testing going on for an extended period, likely a huge upload is occurring. If that is happening, it is highly possible that in the low/medium/high ECM limits phase, that the factors will not "show up" on your screen for up to 30 minutes even though the DB has found them. I think this has to do with the fact that the primality testing of the huge amount of factors and the low/medium/high ECM limits phases all generally occur on the same "subset" of workers. This may be because the DB needs to primality prove the factors that it has found and that proofing was falling in the queue behind the huge uploading of factors that was being done. The very high limits phase appears to occur on a separate set of workers and so did not have that particular issue. The DB seem to "intuivately" know that smaller factors were in the works for primality proofing so that it did not otherwise duplicate work by doing the very high limits phase. (If I'm right on this, that's a very ingenious internal check to avoid duplicating work that is still "internally in progress". Nice work there Syd!)

Remedy for searchers:
If more than once, the "ECM very high limits" phase stops after the "Preparing for ECM very high limits" phase for no reason that you can see, then likely the DB has found smaller factors but your are unable to "see" them yet because they may be queued up for primality proofing behind the huge # of factors that are being uploaded. Your only choice is to wait about 30 minutes or so for them to show up.

I did this over a 2-hour time frame, that is waiting 30 minutes, before continuing past the ECM medium limits phase and it worked great, even while the huge upload was continuing. I just couldn't get nearly as much done.

Alas, with the upload now done, everything is peachie now. :smile: Let the Aliquot DB games continue. :grin:


Gary

kar_bon 2009-07-06 06:17

i thought so too, because only worker #4 will proove for primality!

so every PRP found, has to be proven by this one worker.

i noticed, if i use the submitform for reporting factors and uncheck the 'Insert in background' option, the factors will proven instantly! (i've done this only once during that big amount of factors, so perhaps not the normal way).

10metreh 2009-07-06 08:31

Gary, I think I know what the database problem is. Someone is "babysitting" a C166 and someone (the same person?) is doing the same on a C152, so that whenever another job comes in, they immediately stop it and start their number again. The C166 is a quasi-repdigit, I have no idea where the C152 comes from. These numbers are hogging the workers so that a C105 and a C89 are getting little work.

kar_bon 2009-07-06 08:37

what about this:

i noticed, that many (don't know exactly how much) seqs with higher n (upto n=1M) were uploaded in the database. currently for n=915192.

what about a thread where everyone can post the seqs he/she uploaded!?

i also could create a page for that too.

10metreh 2009-07-06 08:51

Syd has been on again (on the 4th), but he seems to have gone again. It is his birthday today, maybe he's gone on holiday?

gd_barnes 2009-07-06 09:29

[quote=10metreh;179888]Gary, I think I know what the database problem is. Someone is "babysitting" a C166 and someone (the same person?) is doing the same on a C152, so that whenever another job comes in, they immediately stop it and start their number again. The C166 is a quasi-repdigit, I have no idea where the C152 comes from. These numbers are hogging the workers so that a C105 and a C89 are getting little work.[/quote]


I don't believe so. Here's why: I've been on there on and off at many different times over the last couple of days while the load of factors was uploading and I've checked for all kinds of chananigans like this. If you're on there as much as I have been, you get a feel pretty fast for who is doing what (and if they are doing it to anyone else. lol) That definitely was not the case here. If I had consistently observed someone stopping mine or other's jobs like that, I would have stopped theirs and complained in a public thread somewhere! There are some that are somewhat tenancious about putting their job at the top but I've observed none that will overtly stop other's jobs consistently. I've suspected that someone has stopped my job(s) only twice after many hours on the DB over 2 weeks and then it was happening virtually every time over several hours on 2 different days while the big pile of factors was uploading. No one is that consistent about stopping someone or everyone else's jobs.

The only choice that I could see was to do all of your sequences to medium limits, walk away for 30 mins., refresh, and continue from there as needed or necessary.

On a related note, there was a pile of factors that came in again about an hour ago and I had the exact same symptoms, but only for 10 mins. while they were uploading. Then it was done and the symptoms went away. There is little doubt in my mind what is happening after having seen it correlate perfectly many hours after Karsten's and Tim's huge load of factors. I'm now nearly 100% confident that these symptoms are a direct result of a large download of factors. The download doesn't have to be extremely huge and take many hours but it does have to be intense; i.e. enough to clog the primality proving workers for more than a minute or two.

There is another symptom that you can spot sooner: After getting no factors on low and medium limits, when you do ECM high limits (not very high), keep hitting refresh every 2-3 seconds. If it comes back as though it has completed in < ~10 seconds, than you know the problem is happening. High limits should never complete so quickly. That is a pattern that I recognized both before and about an hour ago. I'm pretty sure that it means the DB isn't really "taking" the high limits (not very high) request because it has likely already completely factored the number internally -or- it is getting confused because you are giving it a request on a number that it already has a smaller factor for but that is not showing up on your screen.

Here is what I might suggest: If you are going to upload factors that you think will take the DB more than about 2 minutes to fully primality check, please post it publicly. Perhaps a thread for "large factor downloads", say, for ~10 or more full sequences, would be a good idea.

Having a thread that shows any sequence uploaded (even 1-2 at a time), I think would be redundant with the simple reservations and statuses threads. Kind of like I did earlier here with 147960, I would just suggest that people show it in the status thread. If they are going to keep searching the sequence, then also show it in the reservations thread. Otherwise, as you indicated to me in the reservations thread, there's no reason to reserve it if you aren't going to take it above it's previously reported status.

BTW, I have noticed that the quasi-repeat digit folks (or one person?) are some of the more tenacious about putting their jobs at the tops of the queues so it doesn't surprise me that you got that impression in this particular case. The only time I'm more tenacious about my own is if mine is < 100 digits and theirs is > 110 digits. At 400 curves vs. 4100 curves (very high limit total), I won't wait for that! lol I'll keep kicking mine to the top until they give up or mine complete, which usually takes < 15 mins. I then let them have at it for a while after mine are done. I noticed that they are now a little better about letting smaller jobs complete. It only makes sense for them to wait when they have a job that would take likely 4-6 hours or more, even if no one got in front of them, and yours will only take 15 mins.

Edit: I'm out of time to post my load of updated statuses. I have a few new reservations but am done with a lot more than what I've started as new. I'll post them later on Monday.


Gary

Mini-Geek 2009-07-06 11:30

[quote=gd_barnes;179868]The DB seem to "intuivately" know that smaller factors were in the works for primality proofing so that it did not otherwise duplicate work by doing the very high limits phase. (If I'm right on this, that's a very ingenious internal check to avoid duplicating work that is still "internally in progress". Nice work there Syd!)[/quote]
Depending on just how small the factors are, it's possible that what's actually going on is that the VHL prep is finding the factors. In case you haven't noticed, the VHL prep means that a worker is running some P-1 and P+1 on the number to try to discover factors before it's sent to all the workers for ECM.

Greebley 2009-07-06 17:52

[quote=kar_bon;179889]what about this:

i noticed, that many (don't know exactly how much) seqs with higher n (upto n=1M) were uploaded in the database. currently for n=915192.

what about a thread where everyone can post the seqs he/she uploaded!?

i also could create a page for that too.[/quote]

This is most likely me. I have run around around 240 sequences to 80 digits and then submitted with the aliqueit -s. I wouldn't expect this to use a lot of resources because it takes a long time to get to 80 digits - I am starting with the ones of 10 or so digits and the average seems beteween 1/2 and 1 hour while the -s which waits for a reply and takes 20-30 seconds. I can post the list but they are nearly random - mostly between 400k and 1 million.

It is likely though if it stops between 80 and 85 digits and is a bigger number but less than 1 million, it is my program that is submitting it.

If we could get the sequences from Wolfgang's Creyaufmueller's C80 file and load them into the db, it would save a lot of time actually. There is a whole lot of them however. My program would likely take 100 days to complete the list by itself.

kar_bon 2009-07-06 19:01

[b][size=+2]It's DONE![/size][/b]

i've uploaded the Summary-page for the range 100k<n<200k with [b]all[/b] open sequences!

Currently there're 956 seqs open and 9 errors so far in the Factoring Database (PM sent to Syd).

I've included some small Statistics, too.
I wrote a script to generate those data from the html-page, but if anyone could check this please! thanks!

so now we got a basement for this range!

thanks Mini-Geek for uploading most of the seqs and Frank for the datatable!

10metreh 2009-07-06 19:07

[quote=kar_bon;179949]Currently there're 956 seqs open[/quote]

There should be 957 - the count on Wolfgang's page is 961, and we've found 4 terminating sequences. Has a merge gone unnoticed?

kar_bon 2009-07-06 19:32

Frank's page [url=http://www.frontiernet.net/~aliquot/100000.txt]here[/url] contains the same amount of seqs, so somewhere one seq. must be lost from Wolfgang's page to Frank's page or we missed an ended one!

kar_bon 2009-07-06 19:57

this is the error:

Wolfgang counted on this [url=http://www.aliquot.de/tabellen/tables1.htm]page[/url] in the range for 108000 - 110000 21 open seqs, but there're only 20!

Greebley 2009-07-06 23:44

Wolfgang's page may have some errors. I found at least one and sent him a message - he agreed but though the number small. I hope to try to see if there are others, but I did get discrepencies in my counts.

For example, I believe 10212 is not in his C80 list. My guess is they are not all marked with the lowest sequence possible - so the sequence count is right but the values may not be lowest. That was the case for the one error I found.

10metreh 2009-07-07 06:49

[quote=kar_bon;179955]this is the error:

Wolfgang counted on this [URL="http://www.aliquot.de/tabellen/tables1.htm"]page[/URL] in the range for 108000 - 110000 21 open seqs, but there're only 20![/quote]

Got it!

He's counting side-sequences as open sequences.

kar_bon 2009-07-07 07:35

[QUOTE=10metreh;180011]Got it!

He's counting side-sequences as open sequences.[/QUOTE]

i think not!

see (the link i gave) at the range 168k-170k: there's a side-seq, too, but not counted.

i think, he just forgot to update the '21' in the 108k-110k range or just miscounted!

therefore his total is just one counting to high!
so his corrected 960 open seqs and our 4 found makes 956! the number i got!

Greebley 2009-07-09 07:18

Since these were all updated, as a side project I was looking at entries in the 100k range that neither terminated or ended in a cycle or prime in the database. These are mostly turning out to be entries that get above c60 but terminate before c80 or merge as a side sequence. Those that aren't are the broken ones. The non-broken ones are updated in the DB.

This should provide a double-check that no sequence somehow got lost or dropped along the way and isn't on the list. So far the list is looking good.

kar_bon 2009-07-09 10:12

i've just uploaded all 3 Summary-pages with current data and reservations (hope so).

new:
- some references given for every page where the data came from
- sizes/lengths inserted for the range 250000<n<300000

there're currently about 2800 sequences shown on those 3 pages.

(with 'my' other project on [url]www.rieselprime.de[/url] with about 150000 Rieselprimes from 8000 different k-values and about 13M LLRnet-results where is my crown? -> king of data :showoff: :flex:)

10metreh 2009-07-09 10:26

On the summary page it says mataje took 228042 to 100 digits. He actually took 228024 to 100 digits and did nothing to 228042.

kar_bon 2009-07-09 10:45

corrected. thanks!

Andi_HB 2009-07-09 10:45

[quote=kar_bon;180314]i've just uploaded all 3 Summary-pages with current data and reservations (hope so).

new:
- some references given for every page where the data came from
- sizes/lengths inserted for the range 250000<n<300000
[/quote]

This sequences are missing
250044
250344
250410
250452
250460
250500
I have reserved this in the other Thread

Regards Andi_HB

kar_bon 2009-07-09 10:48

they are there.
i've put the reserved seqs on top of those pages, because everyone can see them first.

Andi_HB 2009-07-09 11:04

Oh - that`s new :surrender

kar_bon 2009-07-16 21:41

[b]News:[/b]

i've inserted the Aliquot related pages in the menu of my [url=www.rieselprime.de]RieselPrimeDatabase[/url] with some more to come.

so far i've crosschecked/inserted all data for 200k<n<250k so far.

i will compile tomorrow a page for the 300k<n<400k range to show the terminated seqs Wolfgang posted,
so the Stats for that range should be uptodate than!

kar_bon 2009-07-21 16:04

newer news:

i've uploaded [b]all[/b] sequences from W.Crayaufmueller upto k=1000000 into the Summary pages (link above).

Wolfgang gave the count upto 1M for open sequenzes as 9446 (minus some we found).

the stats i made show, there're 'only' 9386 open seqs.

so perhaps there're about 60 missings as Greebley mentioned.

ToDo:

- mark the seqs with error in database Greebley reported
- find those missing seqs -> PM me, if you know some

Greebley 2009-07-21 16:31

There is the 'Sterns List' - Sequences not in C80 (unless it was updated since I downloaded), that were part of WolfCrey's download and are up to 100 digits in the DB. These would already be marked as at 100 digits from Stern on the summary page. They weren't all up to 100 digits in the db until recently.

I am unsure why they were not in C80 - they were all between 10,212 and 50 thousand.

I suspect the 9446 count did count these. Not sure if you also counted them or not. There was something like 15 or so in this group.

mataje 2009-07-23 16:54

1 Attachment(s)
First, [URL="http://factordb.com/search.php?query=&se=1&aq=493410&action=last&fr=&to="]493410 [/URL]merge. Now, my results:
[URL="http://factordb.com/search.php?query=&se=1&aq=381600&action=last&fr=&to="]381600[/URL]: line 892, size 100, 2^2*7
[URL="http://factordb.com/search.php?query=&se=1&aq=394200&action=last&fr=&to="]394200[/URL]: line 1070, size 107, 2^2*7
[URL="http://factordb.com/search.php?query=&se=1&aq=419448&action=last&fr=&to="]419448[/URL]: line 1298, size 102, 2^5*3*7
[URL="http://factordb.com/search.php?query=&se=1&aq=421104&action=last&fr=&to="]421104[/URL]: line 3695, size 100, 2^4*5*7*31
[URL="http://factordb.com/search.php?query=&se=1&aq=443016&action=last&fr=&to="]443016[/URL]: line 1738, size 100, 2^2*7
[URL="http://factordb.com/search.php?query=&se=1&aq=451920&action=last&fr=&to="]451920[/URL]: line 603, size 100, 2^3*3*5^2
[URL="http://factordb.com/search.php?query=&se=1&aq=453184&action=last&fr=&to="]453184[/URL]: line 996, size 100, 2^3*3
[URL="http://factordb.com/search.php?query=&se=1&aq=465150&action=last&fr=&to="]465150 [/URL]merge
[URL="http://factordb.com/search.php?query=&se=1&aq=479632&action=last&fr=&to="]479632[/URL]: line 1221, size 101, 2^9*3*11*31
[URL="http://factordb.com/search.php?query=&se=1&aq=490332&action=last&fr=&to="]490332[/URL]: line 752, size 107, 2^3*3^2*7
[URL="http://factordb.com/search.php?query=&se=1&aq=499560&action=last&fr=&to="]499560[/URL]: line 753 (line 650 is bad!), size 100, 2^5*3*7

kar_bon 2009-07-23 17:17

[QUOTE=mataje;182367][URL="http://factordb.com/search.php?query=&se=1&aq=394200&action=last&fr=&to="]394200[/URL]: line 1070, size 107, 2^2*7
[/QUOTE]

this was done by me! :hello:

10metreh 2009-07-23 17:34

[quote=kar_bon;182093]the stats i made show, there're 'only' 9386 open seqs.[/quote]

The "missing" sequences are mostly in the 600-700k range (38 of them). This was going to be our next subproject, but the large number of missing sequences is drawing me towards 500-600k, as there is only one sequence missing there. 400-500k and 700-900k are all accounted for, but taking the former as our subproject could cause a conflict with Wieb Bosma. There is one sequence missing in the 900-1000k range.

kar_bon 2009-07-23 18:12

[QUOTE=10metreh;182380]The "missing" sequences are mostly in the 600-700k range (38 of them). This was going to be our next subproject, but the large number of missing sequences is drawing me towards 500-600k, as there is only one sequence missing there. 400-500k and 700-900k are all accounted for, but taking the former as our subproject could cause a conflict with Wieb Bosma. There is one sequence missing in the 900-1000k range.[/QUOTE]

which one is missing in 900k-1000k? (need input to complete the summary!)

yesterday i checked 800k-1000k with the latest C9C30-file from Wolfgang and all is online like in the database.
today i updated the 600k-range and found those missing seqs too (mostly in 690k-700k).

have to check against the database and mark broken seqs.

perhaps another choice to check against some other files (C80 but out of date, from 2003!).


All times are UTC. The time now is 19:37.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.