mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Factoring (https://www.mersenneforum.org/forumdisplay.php?f=19)
-   -   Home Primes, Reloaded (https://www.mersenneforum.org/showthread.php?t=12530)

kar_bon 2009-10-01 20:37

Home Primes, Reloaded
 
There was a thread [url=http://www.mersenneforum.org/showthread.php?t=9668]here[/url] for Home Primes of Base 10 (most popular).

The results are collected at [url=http://www.mersennewiki.org/index.php/Category:Home_primes]MersenneWiki[/url] but those pages are hard to read, especially for other bases.

With M.Tervooren's [url=http://factordb.com/search.php]FactoringDatabase[/url] it's easy as can be to hold the results (see 'Sequences' and choose the type of sequence).

The last days I've compiled Summary pages (like that ones for the Aliquot Sequences) at [url=www.rieselprime.de]RieselPrimeDatabase[/url] under the new menu 'Interests'.

For now i've only done pages for Bases 2, 6 and 10, because they are most primeness or popular.

For Base 10 I have to do some more work (sizes only upto n~1300) and including more terminated n's with more data from MersenneWiki.

So to all, who will work on these sequences, please feel free to reserve and post your results here. I will update that pages from time to time.

I'm just extending the range for Base 6 to about n=10000 and factoring some of the lower Base 2 composites.

Happy hunting!

Mini-Geek 2009-10-01 22:08

I extended [URL="http://factordb.com/search.php?se=10&aq=1370"]1370[/URL] and [URL="http://factordb.com/search.php?se=10&aq=10014"]10014[/URL] a few lines.

apocalypse 2009-10-02 00:15

Thanks!

I was going to ask about the Home Primes search status once I got my new computer up and running (parts just arrived today). This is perfect timing for me! :smile: Now I've already pushed HP9942 a few iterations while typing my reply...

kar_bon 2009-10-02 00:35

HP[sub]10[/sub]9942 is already prime (P89) at index 49!

apocalypse 2009-10-02 03:23

Sorry. I meant 9922.

henryzz 2009-10-02 06:19

have we got everything from mersennewiki in the database?

apocalypse 2009-10-02 12:43

I've taken HP[SUB]10[/SUB]9922 to n=59. I've also taken HP[SUB]10[/SUB]9900 to n=58.

rogue 2009-10-02 18:43

Does anyone have a script that can recursively test a home prime sequence until a prime is found? I could probably write my own, but this would save me the effort.

BTW, found a home prime for 3393.

10metreh 2009-10-02 19:04

[QUOTE=rogue;191723]Does anyone have a script that can recursively test a home prime sequence until a prime is found? I could probably write my own, but this would save me the effort.

BTW, found a home prime for 3393.[/QUOTE]

I suppose one could hack aliqueit to make it calculate home primes instead of aliquot sequences...

Mini-Geek 2009-10-02 19:09

[quote=rogue;191723]Does anyone have a script that can recursively test a home prime sequence until a prime is found? I could probably write my own, but this would save me the effort.[/quote]
I don't have or know of one. If nobody has one, it'd be easier to modify aliqueit to compute the next line in the sequence as a home prime (i.e. based on the concatenation of prime factors in base x, x could be fixed at 10 or other) instead of as an aliquot sequence (i.e. based on the sum of all factors) than to write a whole new program.
[quote=rogue;191723] BTW, found a home prime for 3393.[/quote]
Congrats. :smile:

rogue 2009-10-02 19:36

[QUOTE=Mini-Geek;191725]I don't have or know of one. If nobody has one, it'd be easier to modify aliqueit to compute the next line in the sequence as a home prime (i.e. based on the concatenation of prime factors in base x, x could be fixed at 10 or other) instead of as an aliquot sequence (i.e. based on the sum of all factors) than to write a whole new program.[/QUOTE]

I will leave that hacking to someone else because I'm not a big fan of perl.

I was lucky enough to find a second home prime today, although I don't recall which one it was. This is the first time I've actually used the factor database. I would make a few changes, but other than that, it saves me a bit of work.

Mini-Geek 2009-10-02 19:40

[quote=rogue;191726]I will leave that hacking to someone else because I'm not a big fan of perl.[/quote]
aliqueit is written in C++, not Perl. I'd do the modification myself, (or at least try...my lack of C++ experience could be a problem there) but I can't get GMP included properly so it can compile. :sad: Anyone know how to get it all to work? (I'm on Windows, Cygwin is available if needed)

RichD 2009-10-02 21:08

C++ & ECM
 
[QUOTE=Mini-Geek;191727]aliqueit is written in C++, not Perl. I'd do the modification myself, (or at least try...my lack of C++ experience could be a problem there) but I can't get GMP included properly so it can compile. :sad: Anyone know how to get it all to work?[/QUOTE]

The original factorization_worker used execlp in the program.

rogue 2009-10-03 00:45

8608 bites the dust. That's three for me today.

Mini-Geek 2009-10-03 12:13

[quote=RichD;191732]The original factorization_worker used execlp in the program.[/quote]
First off, aliqueit is completely different from either factorization worker script. Second, it needs GMP to be able to handle large numbers, not GMP-ECM. Like the worker scripts, it calls GMP-ECM externally. (which is completely separate from what I'm trying to do)

Taking 1315, 1316, and 1328 to a c100 (at least).

RichD 2009-10-03 13:30

C++ & GMP
 
[QUOTE=Mini-Geek;191784]First off, aliqueit is completely different from either factorization worker script. Second, it needs GMP to be able to handle large numbers, not GMP-ECM. Like the worker scripts, it calls GMP-ECM externally. (which is completely separate from what I'm trying to do)[/QUOTE]

Sorry, I thought that might be an easier implementation.

For GMP you need to include GMP headers in the source.
#include <gmp.h>

During the linking process you need to include the GMP library.
gcc hp.cpp -O3 -lgmp hp.out
Notice the -lgmp above.

Anything more specific I won't be of help, maybe others.

mataje 2009-10-05 12:20

Home prime base 10: 2104 finished.

gd_barnes 2009-10-05 13:05

I extended 10 of the base 10 home prime sequences. I found a prime for 6069. I'm not reserving these. Final statuses:

6069, i=47, sz 89 prime

2114, i=57, sz 94, C90
6785, i=50, sz 89, C83
6812, i=48, sz 84, C76
8890, i=38, sz 80, C75
8949, i=53, sz 88, C78
9773, i=55, sz 95, C75
10031, i=60, sz 111, C99
10052, i=51, sz 103, C70
10094, i=60, sz 107, C102


I chose these because they all had a remaining factor of C<70 before I started on them. There are still 10 more of them in the DB. I'll finish them up later today if no one else has done so.


Gary

mataje 2009-10-05 17:22

Home prime base 10: 1855 finished.
I have seen that 1539 is finished too.

Mini-Geek 2009-10-05 18:27

[quote=Mini-Geek;191784]Taking 1315, 1316, and 1328 to a c100 (at least).[/quote]
Done, all have a moderate amount of ECM (several Quick ECMs at least), but not a full battery.

mataje 2009-10-05 20:25

Home prime base 10: 2480 and 2484 finished.
I have seen that 2248 is finished too.

mataje 2009-10-05 22:45

Another one: 3102

Mini-Geek 2009-10-05 22:51

[URL="http://factordb.com/search.php?se=10&aq=5912"]5912[/URL] terminated after a single Quick ECM (clicked by me :razz:).
Taking 6063, 6078, and 6084 until I get tired of them. :smile:
Edit: [URL="http://factordb.com/search.php?se=10&aq=6078"]6078[/URL] terminated.

kar_bon 2009-10-05 23:00

hey, the last days great efforts were done in the HP 10 range!

i've updated the summary just a minute ago: all terminations now in.

as from my script to compare the database with the summary i saw that some steps for many values were done. but i can't update this every day.
i try to keep the terminations current!

mataje 2009-10-06 11:54

Base 10: 3336 finished.

mataje 2009-10-06 13:41

3 more finished: 3492, 3498 and 3586.

Mini-Geek 2009-10-06 14:23

[quote=Mini-Geek;191970]Taking 6063, 6078, and 6084 until I get tired of them. :smile:
Edit: [URL="http://factordb.com/search.php?se=10&aq=6078"]6078[/URL] terminated.[/quote]
Done with 6063 and 6084, both on big composites that passed full ECM (c110 and c109, respectively).

mataje 2009-10-06 19:07

More finished: 3905, 3970, 4012 and 4020.

smh 2009-10-06 21:19

9513 finishes

mataje 2009-10-06 22:48

4109 and 4240 finished. The same with 4318.

gd_barnes 2009-10-07 01:29

I extended the 10 additional base 10 sequences that had a composite factor of C<70:

2802, i=60, sz 104, C94
3685, i=49, sz 78, C71
3852, i=46, sz 81, C73
6336, i=44, sz 79, C75
6680, i=45, sz 86, C74
9080, i=46, sz 100, C83
9975, i=47, sz 94, C85
9983, i=53, sz 105, C89
10098, i=49, sz 93, C70
10132, i=51, sz 85, C82

As of the time I finished this about 12 hours ago, all base 10 sequences then had composite factors of C>=70. That was my objective. :smile: Over the next month, I'll probably work my way up the list to C>=80 or higher unless someone beats me to it.

Edit: Karsten, it looks like you had already updated your pages on Oct. 6th for this effort. I promise I was the one who brought these up to those current level. :-)


Gary

gd_barnes 2009-10-07 02:40

It looked like sequences > 10K needed a little work so I did the ones with composite factors of C70-C75 to get their remaining factors up to C>75. In the process, I found a prime on one of them. Statuses:

10234, i=41, sz 79, prime! :smile:

10045, i=41, sz 87, C86
10050, i=56, sz 97, C78
10052, i=56, sz 118, C109
10058, i=56, sz 95, C94
10098, i=53, sz 105, C96
10136, i=59, sz 98, C94
10200, i=49, sz 102, C93
10203, i=55, sz 103, C95
10208, i=48; sz 81, C81
10218, i=50, sz 87, C82


Although I didn't advance the following sequence, I did find a somewhat large factor for it:

1681, i=57, sz 107, C80

Edit: I also decided to get rid of the final two C70's in the DB. Here are the statuses of them now:

8592, i=45, sz 86, C81
8960, i=41, sz 82, C76


All sequences now have remaining factors of C>70. Getting rid of the C71's should take a little while. :-)


Gary

rogue 2009-10-07 03:03

7072 and 7403 have bitten the dust.

paleseptember 2009-10-07 04:01

[tex]HP_{10}27684[/tex] is prime. Was trying (not-so) random starting points :D

mdettweiler 2009-10-07 05:56

I've advanced [url=http://factordb.com/search.php?se=10&aq=6138&action=last20&fr=&to=]6138[/url] a bit, to i=52, size 93, C93.

richs 2009-10-07 06:19

8960 to i54, size 104, C101

kar_bon 2009-10-07 06:35

[QUOTE=mdettweiler;192067]I've advanced [url=http://factordb.com/search.php?se=10&aq=6138&action=last20&fr=&to=]6138[/url] a bit, to i=52, size 93, C93.[/QUOTE]

[b]recommended to all users:[/b]

i saw the index 52 was a pure C93, so i tested it with the QuickECM-option of the FactoringDatabase.

the first run turns it in P6*C87

the second run splitted the C87 into P20*P67!

so please, try the great QuickECM-option before giving up! :grin:

now it's at index 56 with a C96 (no more factors with QuickECM > next factor should be greater 30 digits)

gd_barnes 2009-10-07 07:38

It's interesting that you mention that Karsten. ~80% of the work that I did above was solely with quick ECM. I had to use msieve to initially break a majority of the C62-C75 but after that, quick ECM usually took it up many indexes with only a couple of msieve runs thrown in.

Hint to all: I always do quick ECM at least 5 times on each # before giving up or using msieve. Many times it will factor on the 3rd/4th/5th try. I can even remember one day I had my 2 cores that I usually run msieve on running on something else and I didn't want to slow it down by running a 3rd program on the machine. So I tried doing quick ECM on a C69 something like 10-15 times and wouldn't you know...it factored on the final try. I remember I was going to give up after 2 more attempts.

Quick ECM is a great tool but it is a little quirky. Don't give up after trying it once or twice. I suspect there is some sort of built in curve randomizer so that it tries different curves each time you hit it. So you may try it 10 times with no luck but alas, it will factor it for you on the 11th try.

BTW, in response to your final sentence about the next factor "should" be at least 30 digits. That's not necessarily true. I've hit quick ECM 10 times and it still didn't end up finding what turned out to be a P28 or P29. On the other hand, I've had it find a P34 or P35 on the first try. Like I said, it can be a little quirky. Several weeks ago, I remember it even had a hard time factoring some C21s thru C23s. I saw the issue brought up in another thread and apparently it got fixed.


Gary

kar_bon 2009-10-07 08:16

the 30 digits is only my observation. i got a P28 or P29 left, too. or QuickECM found some P32 as well.

instead of msieve i use yafu (version 1.12 with an ini-file and 4 threads on my Quad) because it's way faster for upto 90 digits. because yafu don't use 100% processtime when running (i think because of saving the relations found) other programs will continue their work (at about 5-10% CPU-use) although yafu can end the job quite quick!
so no need to shut down other programs and start yafu.
the HomePrimes are growing fast in digits so there is not much work to do manually.

at lower index i need 3 (sometimes 4 when a number is not fully factored) mouse-clicks for an index:
- open a seq in a new window: the last unfactored part will be shown below "Sequence ended: not all factors known"
a) click (1) on the factor
b) click (2) "Quick ECM" -> factors shown
c) click (3) on the seq-number below "Additional info"

begin at a) again.
at higher index you habe to click more as mentioned.

it's the easiest way until there's no special program like aliqueit

gd_barnes 2009-10-07 09:29

That's exactly the way I've been advancing many of the sequences, sometimes 10 indexes or more starting from a C70. I'll have to try yafu instead.

In the mean time, I tried a different way of looking at your page. I decided to attempt to factor using only fast ECM (or msieve on C<=75 that would not fast ECM) on all of the sequences where there was either a 0 or 1 digit difference between the size and the remaining factor. Kind of like what you did with Max's sequence when you saw there was no difference. I figured those had a better chance than normal of not having the fast ECM previously used. There were probably about 30-35 of them and I was able to advance 12 sequences. A few were only advanced 2-3 indexes but several were able to be advanced by 5-10 indexes. I was disappointed not to find any primes. Here they are:

2138, i=59, sz 106, C98
2147, i=64, sz 104, C95
2618, i=46, sz 95, C85
3192, i=51, sz 92, C79
3215, i=54, sz 102, C101*
3867, i=53, sz 99, C97
5368, i=56, sz 103, C93
5632, i=44, sz 93, C88
7210, i=59, sz 111, C82
7554, i=51, sz 100, C97
8729, i=47, sz 87, C86*
9529, i=54, sz 88, C81

* - Advanced from a difference of 1 that easily factored to a different difference of 1 that would not easily factor.

There are a TON of C71's and C72's in the DB. (I got rid of all C<=70.) :smile: I would guess that 20-30% of those will fast ECM. If not, then a quick msieve/yafu to factor them followed by fast ECM to advance them several more indexes will be quite effective. I think we're still a long way away from not being able to easily use fast ECM for home prime base 10 sequences <= 10200.


Gary

gd_barnes 2009-10-07 09:48

I found two C<70 that popped up in the DB so I had to get rid of them. One worked its way to a prime! Here they are:

7343, i=44, sz 78, prime :smile:
7615, i=51, sz 95, C85

mataje 2009-10-07 11:33

I am doing C<80 using Quick ECM and Yafu. N<4500 done.

RichD 2009-10-07 12:31

c74
 
At first I was randomly picking a few HP(10)s to factor. Then I started following this thread so I won't step on other's toes. I began around c74's. Maybe doing a c73 & c75 here or there. It was so easy at first I popped 5-6 the first night. Didn't think it was worth reporting. Then it slowed down. Last night will be my last night for about 5 days before I can do anymore. I probably did 10-12 but I have no record of which ones.

Going forward I may try to keep a list of mine.

RichD.

richs 2009-10-07 13:32

9937 to i58, size 122, C114

mataje 2009-10-07 13:39

4776 and 4891 bites the dust.

smh 2009-10-07 13:43

and so does 8049

Mini-Geek 2009-10-07 13:53

Remember that Quick ECM isn't 'free curves', it's just curves being run on someone else's computer. Running it a dozen times is wasting a lot of CPU time, even if it will sometimes find factors above 30 digits.
The sort of curves it runs are inefficient for finding 30+ digit factors, they're more suited to the 15-20 range; e.g. the final curve (with [URL="http://www.mersenneforum.org/showpost.php?p=188677&postcount=563"]Quick ECM's params[/URL], which has B1=103971) has: (all the following results were found on my PC, from GMP-ECM with -v, on a c79)
[code]Expected time to find a factor of n digits:
20 25 30 35
14.84s 1.97m 22.92m 5.91h[/code]While the B1 sizes optimized for each of those digit levels (B1=11000, 50000, 250000, and 1000000, for 20, 25, 30, and 35 digit factors, respectively) is: (this was copy/pasted together, not originally a single output)
[code]Expected time to find a factor of n digits:
20 25 30 35
10.41s 2.18m 18.37m 2.65h[/code]So clearly for any particular (medium-to-large) size that you're looking for, Quick ECM isn't very efficient. Especially for 30-35 digit factors. (remember that I took the final curve, when really it's 3 parallel runnings of 40 curves going from very small with a practically 0 chance of finding a 30-35 digit factor to the one above)
One thing I wasn't expecting was the result for 25 digits. Apparently having a B1 of about 100000 produces a lower expected time to find a 25-digit factor than with a B1 of 50000! B1=103971 took 1172 ms and expects 101 runs to find a 25-digit factor, for a total of 118.372 seconds, while 50000 took 594 ms and expects 214 runs, or 127.116 seconds. I thought the standard 11000, 50000, ... B1s were optimized to be the lowest times for their digit levels? Seems that it's not, at least for my computer.

Now, if the computer it's running on (the DB server?) will idle otherwise, I say go ahead! :grin: But if it's getting other work done, I just hope you all realize that you're slowing that down in an inefficient way.
[quote=gd_barnes;192075] Quick ECM is a great tool but it is a little quirky. Don't give up after trying it once or twice. I suspect there is some sort of built in curve randomizer so that it tries different curves each time you hit it. So you may try it 10 times with no luck but alas, it will factor it for you on the 11th try.[/quote]
Of course. :smile: It just calls GMP-ECM (a.k.a. ecm.exe, the most widely-used ECM app), which uses a random sigma on each curve.

henryzz 2009-10-07 15:42

[quote=Mini-Geek;192099]One thing I wasn't expecting was the result for 25 digits. Apparently having a B1 of about 100000 produces a lower expected time to find a 25-digit factor than with a B1 of 50000! B1=103971 took 1172 ms and expects 101 runs to find a 25-digit factor, for a total of 118.372 seconds, while 50000 took 594 ms and expects 214 runs, or 127.116 seconds. I thought the standard 11000, 50000, ... B1s were optimized to be the lowest times for their digit levels? Seems that it's not, at least for my computer.[/quote]
those values were optimal at one point but improvements in ecm and in computers have changed what values are optimal
check this [URL="http://www.mersenneforum.org/showthread.php?p=178923&highlight=curves#post178923"]post[/URL] out
i might at somepoint write a program that will parse logs and find optimal values for higher digit levels

mdettweiler 2009-10-07 16:23

[quote=kar_bon;192072][B]recommended to all users:[/B]

i saw the index 52 was a pure C93, so i tested it with the QuickECM-option of the FactoringDatabase.

the first run turns it in P6*C87

the second run splitted the C87 into P20*P67!

so please, try the great QuickECM-option before giving up! :grin:

now it's at index 56 with a C96 (no more factors with QuickECM > next factor should be greater 30 digits)[/quote]
Okay, I'll remember that next time. I'm not sure why I didn't do at least quick ECM for that particular one. :rolleyes:

Meanwhile, extended [url=http://factordb.com/search.php?se=10&aq=8739&action=last20&fr=&to=]8739[/url] to i=50, size 83, on a C79 that wouldn't break after ~5 Quick ECMs.

Same story for [url=http://factordb.com/search.php?se=10&aq=8873&action=last20&fr=&to=]8873[/url], now at i=48, size 88, C84.

richs 2009-10-07 16:41

9027 terminates at step 48

henryzz 2009-10-07 18:16

[quote=gd_barnes;192081]There are a TON of C71's and C72's in the DB. (I got rid of all C<=70.) :smile: I would guess that 20-30% of those will fast ECM. If not, then a quick msieve/yafu to factor them followed by fast ECM to advance them several more indexes will be quite effective. I think we're still a long way away from not being able to easily use fast ECM for home prime base 10 sequences <= 10200.[/quote]
i just did all c71-73
most factored with fast ECM not even 20-30% didn't
continuing onward and upward

smh 2009-10-07 18:24

7188 and 7282 Terminate

gd_barnes 2009-10-07 20:37

[quote=henryzz;192127]i just did all c71-73
most factored with fast ECM not even 20-30% didn't
continuing onward and upward[/quote]

Holy crap. I had to check for myself because that's a ton of pointing and clicking! I find it kind of addictive breaking up those bad boys. It's kind of fun isn't it? lol I see now that there are 5 sequences that are C<=73, which I'm sure people are in the middle of working on so I won't mess with them.

Now here is what I expect you to do: Report every sequence here that you did so Karsten can update his page. (lol)

Seriously, my feeling is that we should work on getting them all up to a "hard" C>=80, that is C>=80 that won't fast ECM.

Tim,

That's a good point on doing too many fast ECM's using up Syd's (or someone's) CPU cycles. I just remember from before when there were a lot of workers so there was apparently quite a bit available, which there isn't now. In the future, I'll limit the fast ECM's to 3-5 attempts. That's usually my limit but sometimes I lose count and end up doing 10 attempts.

That was a good explanation on the curve randomizer. Now I kind of understand why fast ECM appears a little quirky at times.


Gary

gd_barnes 2009-10-07 20:40

[quote=henryzz;192127]i just did all c71-73
most factored with fast ECM not even 20-30% didn't
continuing onward and upward[/quote]

Since you did all those, did you find any primes? I would think you would have found several.

One more thing: On the ones you did, did you continue fast ECMing subsequent indexes until you hit a tough one? If not, I take back what I said about it being a ton of work. lol

mataje 2009-10-07 21:17

I have seen that 6212 is finished.
For N<6212 c<80 done.

richs 2009-10-07 21:22

Instead of using someone else's CPU cycles, it is very easy to factor up to C85 composites using Dario Alpern's ECM website:

[url]http://www.alpertron.com.ar/ECM.HTM[/url]

Cut and paste the composite number into the top bar on this site and press return. When you get a factor, cut and paste it back into the factoring db. Very easy....

mataje 2009-10-07 21:48

6256 is finished. :smile:

kar_bon 2009-10-07 23:29

the lowest composites of all open HP10 in the FactoringDatabase are C75!

see [url]http://factordb.com/search.php?so=1&se=10&5=1&start=2&limit=1000&ew=1[/url] at the bottom.

rogue 2009-10-07 23:34

[QUOTE=richs;192143]Instead of using someone else's CPU cycles, it is very easy to factor up to C85 composites using Dario Alpern's ECM website:

[url]http://www.alpertron.com.ar/ECM.HTM[/url]

Cut and paste the composite number into the top bar on this site and press return. When you get a factor, cut and paste it back into the factoring db. Very easy....[/QUOTE]
[URL="http://www.mersenneforum.org/showthread.php?t=9629"]this[/URL] would be faster than his applet. I believe that msieve will also extract small factors before using mpqs.

BTW, how does one report multiple factors at a time.

kar_bon 2009-10-07 23:36

[QUOTE=rogue;192152]BTW, how does one report multiple factors at a time.[/QUOTE]

insert them in the "Report factor(s)" window, one per line or comma-separated

rogue 2009-10-08 01:03

[QUOTE=kar_bon;192153]insert them in the "Report factor(s)" window, one per line or comma-separated[/QUOTE]

I found out that factordb will take output from my factoring program. I presume it removes non-numerics when parsing. This means that it will accept output such as this:

[code]
3 (prime)
* 3 (prime)
* 13 (prime)
* 19 (prime)
* 5099 (prime)
* 146472107 (prp)
* 20874639816122426922424808819059395515734254643644518586583484290147440406619 (prp)
[/code]

It also takes msieve output such as this:

[code]
p5 factor: 10711
p6 factor: 318503
prp86 factor: 97118834888157129964637503081673516437296046739370628771378517672228613658237113346843
[/code]

Mini-Geek 2009-10-08 02:17

[quote=rogue;192155]I found out that factordb will take output from my factoring program. I presume it removes non-numerics when parsing. This means that it will accept output such as this:

...[/quote]
It's pretty flexible. For the most part, if it's there (and not a single number with line separations in it), it'll find it and enter it.

mdettweiler 2009-10-08 04:40

[quote=gd_barnes;192140]That's a good point on doing too many fast ECM's using up Syd's (or someone's) CPU cycles. I just remember from before when there were a lot of workers so there was apparently quite a bit available, which there isn't now. In the future, I'll limit the fast ECM's to 3-5 attempts. That's usually my limit but sometimes I lose count and end up doing 10 attempts.[/quote]
I belive Syd said the Quick ECM button runs curves on the webserver itself. Syd, if you're reading this thread, is that machine running any other work that the Quick ECM is cutting into, or is it otherwise idle? If it's otherwise idle, then it would actually be better for us to use Quick ECM as much as is needed (though of course not excessively) so as to put that to good use. :smile:

henryzz 2009-10-08 06:16

[quote=gd_barnes;192141]Since you did all those, did you find any primes? I would think you would have found several.

One more thing: On the ones you did, did you continue fast ECMing subsequent indexes until you hit a tough one? If not, I take back what I said about it being a ton of work. lol[/quote]
i have not a clue
the way i do it is select the right side of the page while holding ctrl(otherwise it selects the whole page) and use firefox's open links in selection feature
i then have all the pages for all the composites(can take a few seconds to load 50 tabs)
i then go to the last tab opened and click quick ecm and then straight away unless there was already more than one composite factor before clicking on quick ecm type ctrl-w which closes the tab i then repeat for all tabs

i dont even see half the time whether a number is factored until i referesh the sequnce overview page
i you do this be careful not to run too many quick ecms at once it means no one can access the server until they are finished(i suspect they run for 4 seconds of cpu time not clock time)
no way am i writing what i have done for each sequence:smile:
copying the composites out would be the best way of doing it but it would take a rediculous amount of time in comparison

gd_barnes 2009-10-08 10:25

That's a pretty slick way to do it but you lose all the fun of seeing the factorizations. lol

Why wouldn't you have a clue if you found a prime? Can't you refresh all of your pages after the factorizations?

Regardless, IMHO, you should report the sequences that you find a prime for so that Karsten can at least show those on his page without having to sort them in the DB by primed sequences. That's what everyone has been doing here. Besides, half the fun of doing the sequences is finding a prime for them, especially the big primes.

Karsten, if he's doing that, I see almost no reason to report statuses here unless we find a prime. Would you agree?

gd_barnes 2009-10-08 10:29

[quote=richs;192143]Instead of using someone else's CPU cycles, it is very easy to factor up to C85 composites using Dario Alpern's ECM website:

[URL]http://www.alpertron.com.ar/ECM.HTM[/URL]

Cut and paste the composite number into the top bar on this site and press return. When you get a factor, cut and paste it back into the factoring db. Very easy....[/quote]

No! I've used Alpertron's site extensively in the past. It defeats the purpose of why I'm using fast ECM in Syd's DB...to avoid using my own CPU cycles on a dual core while 2 cores run Aliquot sequences. Alpertron's site uses your own CPU cycles not his. Besides, if I wanted to use my own CPU cycles, I'd use msieve or yafu up to C85.


Gary

kar_bon 2009-10-08 10:36

doing 50 n-values at once and only advancing one index is not worth to fill in a contributor in the summary.

if someone is posting here a terminating sequence, i will include the name/date.

my script will find any termination or any difference in the last index, so no need to post here all n's with one new index.

as for base 6 i think i will show all terminations with primes > 70 digits.
some terminations (see summary) i can't say who has done them so only a '?' for the finder until the finder will post here!

i think i can update the summary once a week, but every day it's too much work!


PS: as i can say, there're 17 new terminations not yet shown in the summary since yesterday!

gd_barnes 2009-10-08 11:24

Speaking of terminations:

7024 is prime at i=46 size 77 :smile:

BTW, all the statuses I've posted here have advanced more than one index; mostly 5-10 indexes and sometimes even 10-15.

mataje 2009-10-08 11:35

4557 finished.

Mini-Geek 2009-10-08 12:00

[quote=richs;192143]Instead of using someone else's CPU cycles, it is very easy to factor up to C85 composites using Dario Alpern's ECM website:

[URL]http://www.alpertron.com.ar/ECM.HTM[/URL]

Cut and paste the composite number into the top bar on this site and press return. When you get a factor, cut and paste it back into the factoring db. Very easy....[/quote]
Easy, yes, but inefficient. GMP-ECM is [I]much[/I] faster (roughly 5 times faster, from a quick test, if it uses the B2 that Alpertron's app does, but normally it uses a larger B1). Like Gary said, the whole point is to not pull your current computer's processing power away. If you are going to do ECM locally anyway, might as well make it as efficient as possible.

mataje 2009-10-08 12:58

Another one finished: 6919. The same with 7049.

mataje 2009-10-08 15:12

c<80 done.
Edit: 8170 finished. And 8575

mataje 2009-10-08 18:42

9141 finished.

richs 2009-10-08 19:36

10302 terminates at index 51 with a P88

richs 2009-10-08 20:33

10380 terminates at index 46 with a P90

richs 2009-10-09 05:10

10424 terminates at index 42 with a P82

mataje 2009-10-09 06:46

6906 terminates.

mataje 2009-10-09 13:14

8562 terminates. And 6877 too.

gd_barnes 2009-10-10 02:12

Well, I was going to post a status of 9 sequences that I added some indexes to early yesterday but with them all having been advanced to C>=84, that would be somewhat pointless.

Nice work everyone getting all of the "low hanging fruit" taken care of. Next target: 90 digits for all of the sequences <= 10K -and- they must survive at least one "Fast ECM" (preferrably 3). :smile:

smh 2009-10-10 07:25

9266 Terminates

henryzz 2009-10-10 07:33

i changed my method a little
1846 terminates

henryzz 2009-10-10 09:06

8193 terminates
9890 terminates

mataje 2009-10-10 13:52

6550 terminates.

apocalypse 2009-10-10 17:33

HP[SUB]10[/SUB]9780 terminates n=57, 109 digits

My first HP! :smile:

mataje 2009-10-10 20:38

c<86 finished.

mataje 2009-10-10 23:00

10158 terminates. 10609 terminates. 10555 terminates. 10616 terminates.

apocalypse 2009-10-11 00:35

HP[SUB]10[/SUB]8457 terminates n=62 P100

RichD 2009-10-11 06:31

5496
 
5496 terminates with a p102

mataje 2009-10-11 11:01

10840 terminates. The same with 10668 and 10755.

mataje 2009-10-11 12:12

7316 terminates.

mataje 2009-10-11 13:59

2581 terminates. 10659 terminates.
BTW: c<87 done for N<10700

mataje 2009-10-11 14:40

3388 terminates.

mataje 2009-10-11 14:56

10764 terminates.

apocalypse 2009-10-11 15:17

HP[SUB]10[/SUB]9961 terminates at n=52 P95

mataje 2009-10-11 19:31

10948 terminates.

mataje 2009-10-11 22:57

10991 terminates.
BTW: c<88 finished.

mdettweiler 2009-10-12 03:40

Just now I've been doing a wee bit of work in the 11000-11050 range, which didn't seem to be either in the database or in Karsten's summary. As such, everything was being calculated afresh, leading of course to a lot of new, but completely non-noteworthy, terminations which were found right off the bat by the DB in its initial calculation of the sequences when I looked them up. I'm guessing those aren't worth reporting because Karsten's scripts will automatically pick them up, right?


All times are UTC. The time now is 20:37.

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