[QUOTE=schickel;282759]OK, I ran a poly search while I was at work today, and since no one else has said anything, I just kicked off the job for this. ETA is maybe 3 days? (It's been a while since I did one this size, so I'll know more for sure tomorrow....)[/QUOTE]Make that ~1.5 days. It should finish by the time I get back from work tomorrow.

The factors are in the DB - have you send the factors Schickel?
New 137 digit composite at line 2789 now.

[QUOTE=Andi_HB;282948]The factors are in the DB - have you send the factors Schickel?
New 137 digit composite at line 2789 now.[/QUOTE]Nope, wasn't me. Guess I'll go downstairs and kill the job that's running.....

[QUOTE=Andi_HB;282948]The factors are in the DB - have you send the factors Schickel?
New 137 digit composite at line 2789 now.[/QUOTE]

Did 3000 curves at B1=10M on the C137. Assuming that others have probably also run some curves, I'll go ahead and start GNFS on it.

I did 1154 curves @ B1=10000000, B2=46842680440 on line 2789 c137

[QUOTE=Andi_HB;282948]The factors are in the DB - have you send the factors Schickel?
New 137 digit composite at line 2789 now.[/QUOTE]

The C137 splits as:
[CODE]
prp65 = 92604252360612149566142017110232157485024433471209368898867397209
prp72 = 360801540857918157146214575783484927806773415487836208586298906480233817[/CODE]

A fact I discovered over at factordb.com while my sieving effort sits at about 95% done.

[QUOTE=schickel;282951]Nope, wasn't me. Guess I'll go downstairs and kill the job that's running.....[/QUOTE]

:doh!:

There is a C133 now at line 2797... I think I'll let our mystery helper continue with it for now. Although the childish part of me wants to throw 200 cores at it and be done within the hour, just so that he/she can be dissapointed for once...

[QUOTE=bsquared;283167]There is a C133 now at line 2797... I think I'll let our mystery helper continue with it for now. Although the childish part of me wants to throw 200 cores at it and be done within the hour, just so that he/she can be dissapointed for once...[/QUOTE]Hee hee, I know that feeling. :tu:

Whew!
 

Man, did we just dodge a bullet!

Line 2795 was this:[code]2795 c155 = 2 * 26540231597764915875748523921286155358287735668816451607271020126944017648528345710110955950716807274798345163208946240501455091962907345828078337693949271[/code] [SIZE="1"](Or c155 = 2 * p155)![/SIZE]

This mystery worker has quite a lot of cores - they're doing c13xs in about a day. Ben, can you throw loads of cores at the c149? It would be nice to get it out of the way quickly anyway.

[QUOTE=10metreh;283298]This mystery worker has quite a lot of cores - they're doing c13xs in about a day. Ben, can you throw loads of cores at the c149? It would be nice to get it out of the way quickly anyway.[/QUOTE]

I can put 48 cores on it, but I can't start until tonight after work (10 hrs from now). I'll start 6 cores on ecm until then.

[QUOTE=bsquared;283317]I'll start 6 cores on ecm until then.[/QUOTE]

Glad I did:

[CODE]>> ecm: 846/2004 curves on C149 input, at B1 = 11M, B2 = gmp-ecm default
ecm: found prp45 factor = 242660030913707404593072169939694042199547793

***factors found***
PRP45 = 242660030913707404593072169939694042199547793
ans = 74335278981037400615182032951788222765485718617195876482109349156724018571270758476761382399263807180313[/CODE]

I've done the following ecm work so far on the C147 of line 2799:
1000 @ 1e6
3000 @ 10e6

1000@44M here so far.

Found 197765000678978787299763869154327257712005970001

Edit: did p40 on the c136, continuing at 11M. Did about 2500@11M, moving on to 44M until someone starts sieving. Edit: Started poly search. Edit: Sieving. Edit: Linear Algebra.

Done. Now a c132 at step 2804. I'm currently running curves at 3M.

[QUOTE=schickel;283206]Man, did we just dodge a bullet!

Line 2795 was this: [SIZE=1]c155 = 2 * p155[/SIZE]
[/QUOTE]

[URL="http://en.wikipedia.org/wiki/MythBusters#Esparto_accident"]MythBusters[/URL] busted the myth that bullets can be dodged. :max:
...and fan cannot dodge certain solid matter.

[SIZE=1]c154 = 2 * p153[/SIZE]

We have been very unlucky when it comes to losing downdrivers.
There is a slightly less than 50% chance that the sequence will pick up a 3 on the next line. :sad:

[QUOTE=10metreh;283699]We have been very unlucky when it comes to losing downdrivers.
There is a slightly less than 50% chance that the sequence will pick up a 3 on the next line. :sad:[/QUOTE]Let's hope that doesn't happen!

Still, since the top, we've had two good runs:

i2642:c175 to i2691:c166, or 50 lines and 9 digits

i2721:c164 to i2804:c152, or 80 lines and 12 digits

Of course, the last one would have been better, if not for 2 runs of 20 lines each where it didn't lose any digits, since it was 2*5 or 2*7.

Maybe we'll get lucky and have another downdriver after ~30 lines or so....

3000 curves done @ 3M on the C152 of line 2805

And another 2k@3M here

running polynomial selection on line 2805

+ 4500@11M
+ 600@110M

Another 3800@11M. These and Ben's count for .44 of p50.

+ 500 @11M
+ 400 @110M

With these we're at ~0.6t50: enough to start sieving, IMO. I'll take on the GNFS job once jrk posts a poly.

[QUOTE=bsquared;283786]+ 500 @11M
+ 400 @110M

With these we're at ~0.6t50: enough to start sieving, IMO. I'll take on the GNFS job once jrk posts a poly.[/QUOTE]

Here you are
[code][color=blue]# sieve with ggnfs siever 14e on alg side from Q=10M to 38M[/color]
# aq4788:2805
n: 35349578600188434935757617497394601840825892193617972555577497260598798875991889870743608819655457195654711342030972300657595238390994384661130562933053
# norm 1.122443e-14 alpha -6.002492 e 4.661e-12 rroots 3
skew: 748189.16
c0: -46774038100489702875504930683561184
c1: 308869238721633468571835584258
c2: 446679487152622914099661
c3: -1625198709211324148
c4: -2576950862160
c5: 255024
Y0: -169184631318470913062170668205
Y1: 52388078770696883
rlim: 20000000
alim: 20000000
lpbr: 29
lpba: 29
mfbr: 58
mfba: 58
rlambda: 2.5
alambda: 2.5
[/code]

Great, thanks! Sieving underway.

jrk,

Did you use CPU (what CPU) or GPU (what nvidia GPU) for poly selection? For how long did you run the selection?

Carlos

ECM stopped. If you like I can throw in some relations if you assign me an sq range.

[QUOTE=akruppa;283820]ECM stopped. If you like I can throw in some relations if you assign me an sq range.[/QUOTE]

Thanks for the offer. I'm currently sieving at about 0.75 MQ per hour, so I'll be done this time tomorrow. 20-30M is currently assigned, I was planning to assign 30-38M tonight. If you want to do a portion of 30-40M and can have it done inside 24 hours from now, go ahead. Just let me know what you take so I don't stomp on your feet.

If you are proceeding that quickly, it's probably not worth the overhead of having me add a few more cpus. We'd save a few hours at most.

[QUOTE=akruppa;283825] We'd save a few hours at most.[/QUOTE]
Kinda what I was thinking too, but thanks again for the offer - you can have the next one if you want :smile:.

Done:

[CODE]prp53 factor: 47741505965405370666781117072020826470386542910028553
prp99 factor: 740437024039492581716955948961933782533368043722223980779075769406246940852840276397465096486276501[/CODE]

But bummer of a next line... :down:

[QUOTE=bsquared;284058]Done:

[CODE]prp53 factor: 47741505965405370666781117072020826470386542910028553
prp99 factor: 740437024039492581716955948961933782533368043722223980779075769406246940852840276397465096486276501[/CODE]

But bummer of a next line... :down:[/QUOTE]

:ouch2:

this moves the sequence to the 2²*3 subproject...

I'll do the c111.

[QUOTE=akruppa;284084]I'll do the c111.[/QUOTE]

Already done... sorry. Didn't see your post till afterwards.

Newbie question. What does it take to lose the 3? What are the chances?

[QUOTE=Prime95;284096]Newbie question. What does it take to lose the 3? What are the chances?[/QUOTE]There is some discussion [URL="http://www.mersenneforum.org/showthread.php?t=16075"]here[/URL], from post #12 through #16(17) about what it takes....unfortunately, it really all boils down to luck of the draw type stuff that depends on how a number factors.

I've ECM'd the c129 to nearly t40. Anyone want the gnfs job?

I did about 1k@11M. If you'd like to do this one, go ahead. If you're otherwise occupied, I can do it.

[QUOTE=akruppa;284131]I did about 1k@11M. If you'd like to do this one, go ahead. If you're otherwise occupied, I can do it.[/QUOTE]

I started it, should have factors within the next couple days.

[QUOTE=schickel;284101]There is some discussion [URL="http://www.mersenneforum.org/showthread.php?t=16075"]here[/URL], from post #12 through #16(17) about what it takes....unfortunately, it really all boils down to luck of the draw type stuff that depends on how a number factors.[/QUOTE]

My amended version of post #12 is:
I think it's something like this:
All primes congruent to 1 mod 3 must be raised to a power n with n != 2 mod 3.
All primes congruent to 2 mod 3 must be raised to an even power

So the easiest way for it to break is to have very few factors besides 2^2*3, and for them all to be 1 mod 3.

Unless I've made a mistake somewhere, this means that the factor of 83 (which is 2 mod 3) in the current line is the first indicator that the guide will not change in the next line. The p23 factor, which is also 2 mod 3, also means it will not change. The remaining c129 (which is 1 mod 3), assuming two distinct, power 1 factors, could either be 1*1 mod 3, or 2*2 mod 3. If it were only up to the c129, then, we'd have roughly a 50/50 chance of it breaking this line.

Animal Farm version: 1 mod 3 gooood, 2 mod 3 baaad. :alex:
:smile:

[QUOTE=bsquared;284138]I started it, should have factors within the next couple days.[/QUOTE]

Done - I'm ECMing the c154 (which is 2 mod 3).

Did 800@1M and a few at 3M on the c104. Anyone want to take that one? Edit: I started it on 4 cpus. If someone can finish this off quickly, please do.

[QUOTE=akruppa;284237]Did 800@1M and a few at 3M on the c104. Anyone want to take that one? Edit: I started it on 4 cpus. If someone can finish this off quickly, please do.[/QUOTE]

Go for it - I might be away from my computer much of today, so run with it for a while if you want.

C150 of 2819 now. I'm doing curves at 1M atm, soon advancing to 3M. 11M. Found a p48, doing the c102 with GNFS.

A c111 of 2821 which I tested to almost p40. Any takers?

[QUOTE=akruppa;284309]A c111 of 2821 which I tested to almost p40. Any takers?[/QUOTE]
I turned a dual-core loose on it. Should have something tomorrow.

Happy New Year!

BTW, feel free, anybody, to run the c111 if I'm too slow...

A few lines added - the 3 is gone and a c130 now awaits. I have my machines working on it, but feel free to take it away...

[QUOTE=EdH;284362]A few lines added - the 3 is gone and a c130 now awaits. I have my machines working on it, but feel free to take it away...[/QUOTE]A nice way to ring in the New Year!

Did 1600@3M. Continuing at 11M. Did 3k@11M.

I've got 6 machines currently running curves@3e6 - probably ~3000 done.

I've also started gnfs and am sieving on three machines using gnfs-lasieve4I13e. The poly is probably not that great; it's turning out just a little over 2:1 relations. I've listed it below.

I might have something tonight, but more likely, tomorrow.

Anyone else is welcome to go ahead and factor this one, though. I would have no hard feelings...

Poly:
[code]
n: 7075503377103996733813023004244260921479211649158454102182330835068475082321737256425610844506878670135904868119017932627153650019
Y0: -33488350220162067167622512
Y1: 58507686675799
c0: -2540911319970342392756569025103867
c1: 618321542130743636360105235
c2: 12135203596104444530201
c3: 3228485668955417
c4: -4583931634
c5: 168
skew: 1716118.22
type: gnfs
[/code]

[QUOTE=EdH;284388]I've got 6 machines currently running curves@3e6 - probably ~3000 done.

I've also started gnfs and am sieving on three machines using gnfs-lasieve4I13e. The poly is probably not that great; it's turning out just a little over 2:1 relations. I've listed it below.

I might have something tonight, but more likely, tomorrow.

Anyone else is welcome to go ahead and factor this one, though. I would have no hard feelings...[/QUOTE]No, you have to use [I]reverse[/I] psychology on the poacher. Make it clear you're making a serious run at it, then they'll jump in just to prove they can do it faster than you....:tu:

{edited to add: maybe take the machines running ECM and add them to the mix -- if it's easy for you to switch them over, that is....}

[QUOTE=schickel;284430]No, you have to use [I]reverse[/I] psychology on the poacher. Make it clear you're making a serious run at it, then they'll jump in just to prove they can do it faster than you....:tu:

{edited to add: maybe take the machines running ECM and add them to the mix -- if it's easy for you to switch them over, that is....}[/QUOTE]
The main machine is currently at 74.3% and I'm probably going to be somewhat tied up tomorrow. I'm hoping I can transfer the relations from the three others tomorrow morning and then maybe it can finish up. Most of the time, the factmsieve.py estimate is low and more sieving is needed. I do have the machine set to upload to the db as soon as the factors are found, but I might not make any note here until later in the day.

Right now I'm hesitant to swap the others over due to trying to figure out the q ranges and then, the concatenation of all the relations files from more than the current three, onto the main machine, might take too much time in the morning.

Of course, I'm hoping all goes well with the main machine. I have it set to carry on past the current line for a few more, submitting each to the db as factored. That should give it a few more tries in case it can't communicate with the db on the current line, providing the next line(s) factor quickly.

Tell me an SQ range (Edit: and sieving paramters) and I'll help sieving a little.

[QUOTE=akruppa;284473]Tell me an SQ range (Edit: and sieving paramters) and I'll help sieving a little.[/QUOTE]
Thanks, but it's running the Linear Algebra step, with an estimated completion around 11:30 AM Eastern U.S. time. With the relations from the other machines, they were at 118.1% this morning.

Found a p41 and we lost the 2^2.

[QUOTE=Batalov;276134]Greg broke it, Greg fixed it! You da man![/QUOTE]
Alex broke it, Alex fixed it! Looks like a pattern. :rolleyes:

[QUOTE=Batalov;284521]Alex broke it, Alex fixed it! Looks like a pattern. :rolleyes:[/QUOTE]

I should check my old test files that I saved up on my other machine because apparently I can get a C101 in Total time:~4 hours according to my last test with my new machine.

+2000 @ 11e6 on c137.

About 4500@11m. Any volunteers for NFS?

[QUOTE=akruppa;284548]About 4500@11m. Any volunteers for NFS?[/QUOTE]
Yes :smile:

[QUOTE=akruppa;284520]Found a p41 and we lost the 2^2.[/QUOTE]Shhhh...don't make too much noise or the other 2 will find us again!

[QUOTE=Batalov;284521]Alex broke it, Alex fixed it! Looks like a pattern. :rolleyes:[/QUOTE]Looks like if you lose the downdriver, you have to run the sequence until you find it again....:rofl:

[QUOTE=Ralf Recker;284565]Yes :smile:[/QUOTE]

Any estimate on the completion time?

[QUOTE=akruppa;285092]Any estimate on the completion time?[/QUOTE]
The factorization should be complete in a few hours.

[CODE]Info:Replaying merge history...
Info:Nrows: 2412653; Ncols: 2412493; Weight: 325141724.
Info:--------------------------------------------------------------------------
Info:Linear algebra
Info:--------------------------------------------------------------------------
Info:Calling Block-Wiedemann...
[/CODE]

GNFS done, P-1 done
 

[CODE]Info:Preparing 23 dependency files
Info:Testing dependency number 0...
Info:Factorization was successful!
Info:282513231304648878202268691132529834582756908313 [prime]
Info:195378052051142638554713151283071968880687961740908823175466158468827553276169042042819791 [prime]
Info:Doing gcds with previously known composite factors
Info:non-trivial factor: 282513231304648878202268691132529834582756908313 [prime]
Info:non-trivial factor: 195378052051142638554713151283071968880687961740908823175466158468827553276169042042819791 [prime]
Info:Now: 2 prime factors, 0 composite factors
Info:Factorization complete
[/CODE]Update (i2840):

[CODE]ecm -pm1 1e9 < 4788-2840-c131
GMP-ECM 6.4 [configured with GMP 5.0.90, --enable-asm-redc] [P-1]
Input number is 24340857327964466671448347911010690599341977754432193301612615831404400353846724664314443468263024608145936303377184861751941023301 (131 digits)
Using B1=1000000000, B2=211252922128080, polynomial x^1, x0=2892386330
Step 1 took 136232ms
Step 2 took 111035ms
********** Factor found in step 2: 22320321079191410444006842537460802229
Found probable prime factor of 38 digits: 22320321079191410444006842537460802229
Probable prime cofactor 1090524515377905728815681621083347896479897110161902243550926350895476187350655021464603295569 has 94 digits
[/CODE]

[QUOTE]Unchecked 2845 155 (show)[/QUOTE]

wow so it's went 5 lines in 1 hour ?

[QUOTE=science_man_88;285175]wow so it's went 5 lines in 1 hour ?[/QUOTE]
Yes.

1000@1e6 on c144.

[CODE]i2843 = 2 · 1997113451...3<155>
i2844 = 2^2 · 2275681 · 436065311457107<15> · 273382045978973731935568292259751669<36> · 1840389861...3<98>[/CODE]:yucky:

[QUOTE=Mini-Geek;285185][CODE]i2843 = 2 · 1997113451...3<155>
i2844 = 2^2 · 2275681 · 436065311457107<15> · 273382045978973731935568292259751669<36> · 1840389861...3<98>[/CODE]:yucky:[/QUOTE]OK, who did it? We'll have to get them to run it until they get it back.....

[QUOTE=schickel;285206]OK, who did it? We'll have to get them to run it until they get it back.....[/QUOTE]

:whistle:

[QUOTE=Ralf Recker;285182]Yes.

1000@1e6 on c144.[/QUOTE]
+2500@11e6

[QUOTE=Ralf Recker;285230]+2500@11e6[/QUOTE]I'll start some @ 43M.

1000 more @ 11e6

[QUOTE=EdH;285286]1000 more @ 11e6[/QUOTE]
+ another 1000 @ 11e6 (total from me, so far, 2000)

[QUOTE=EdH;285286]1000 more @ 11e6[/QUOTE][QUOTE=EdH;285352]+ another 1000 @ 11e6 (total from me, so far, 2000)[/QUOTE]I've done 500 @ 43e6 + 50 @ 110e6.

1000@44M

I'm sorry....that was me and I didn't even notice it:[code]Factor returned by schickel@psln.com:server:v2.1a! a4788.c144 / 60820719675108711017052771789378403947624261197:probable:4915690297241178383115395674859512526367188623932217182205112390254793303831518196606724863836639:Probable finder: schickel@psln.com:Machine_1:v2.0k B1:110000000 sigma: 1481085792[/code]

Found a p31, leaves a c109. Not many curves at 1M done yet, I'll finish the p35 test. Edit: Finished.

1000@1e6 and 2600@3e6 yielded nothing.

Starting NFS.

[QUOTE=akruppa;285423]Starting NFS.[/QUOTE]

You can stop it...

[CODE]GMP-ECM 6.4 [configured with GMP 5.0.90, --enable-asm-redc] [ECM]
Using B1=11000000, B2=35133391030, polynomial Dickson(12), 4 threads
Finished 1090/5000; avg s/curve: stg1 12.83s, stg2 5.526s; runtime: 46s

Run 1090 out of 5000:
Using B1=11000000, B2=35133391030, polynomial Dickson(12), sigma=3985689015
Step 1 took 12900ms
Step 2 took 3853ms
********** Factor found in step 2: 4413824230868329169113745386241314009291154085286949
Found probable prime factor of 52 digits: 4413824230868329169113745386241314009291154085286949
Probable prime cofactor 663064341931092444838761835970780110874338028946807213371 has 57 digits
[/CODE]

46 seconds because I restarted the run to check out the speed of another gmp-ecm build... :blush:

[CODE]-> Found previous job file job2323.txt, will resume work...
-> *** Already completed 1082 curves on this number...
-> *** Will run 3918 more curves.[/CODE]

Did 800@1M on the c127. Continuing at 3M. Edit: did 2000@3M. Edit:2300@11M so far. Continuing.

1000@3e6

+1000@3e6 (2000 total)

1000@1e6
2600@3e6
5000@11e6

500@43e6

Another 800@11M. The total is well above the expected number of curves for a p45, which is more than sufficient ECM testing for a c127. I'll do NFS.

At this point it doesn't really matter, but in case it's of interest, my machines logged an extra 3000@3e6 overnight. (total 5000@3e6)

I should be ashamed of myself!
 

[SPOILER]I forgot to turn the composite off in my server .ini file.....oops!

Factor returned by schickel_psln.com:Machine_1:v2.0k! a4788.c127 / 203963847042504632994125009829787343593285053643:probable[/SPOILER]

:blush: :max:

[SIZE="1"]Sorry, I got a little rushed on the way to work this morning....[/SIZE]

NFS was almost done by the time I saw this but I'm not too worried about it... I'm more interested in seeing the sequence advance then in minimizing CPU time to advance it. Did 2k@3M on the c141, doing 11M now. Did 4000@11M, on to 44M.

1000 @ 1M, 3000 @ 3M and 5000 @ 11M on the c141 done.

1000@11e6

About 3k@44M so far.

1000 more @11e6 (total 2000)

Found a p54 = 113610552357306384103218317618264070899589995889825277
Edit: doing curves at 3M on the c131 now. Edit: 11M.

I have NFS and ECM running against the 2858:c120. It should be finished by tonight, if no one else goes ahead and factors it.

The machine should automatically update the db, so I might not post anything here in a timely manner, if I haven't noticed the completion.

It traded for a c129. Working it...

Doing NFS on the c129.