mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Cunningham Tables (https://www.mersenneforum.org/forumdisplay.php?f=51)
-   -   Contributing to Cunningham Project (https://www.mersenneforum.org/showthread.php?t=24211)

VBCurtis 2021-01-11 23:49

Taking admin 15M, admax 60M, incr 4620, P=4M.

That leaves 10-15M / incr 420 for anyone else to contribute on CADO if they choose.

firejuggler 2021-01-12 03:14

I've got a 1.377e-14 with Msieve, trying to use cado to spin it up.
Record is around 2e-14 so still a bit far

charybdis 2021-01-12 17:20

The best from my run, by some margin, is
[code]n: 28653448550423582065969107478752150888235561427247664110066527668159681894430155521476413904809897034030256344677647748498114768327844226758360594247371333432148627535343266861185808662829021
skew: 61160593.83
c0: -782066809369899992846921498396677009099575965
c1: 129978232145697650128010452670728125268
c2: 1136403038232649046041648047998
c3: -50366553898678575377716
c4: 291334628577375
c5: 2637600
Y0: -6410674956398678515938860432260175622
Y1: 195016503229310813759987119[/code]
Cownoise scores this as 2.023e-14. I tried spinning with EdH's script but didn't get an improvement.

firejuggler 2021-01-12 17:48

That's still close to the record for a C191, so good enough?
afaik

[code]
...

191 2.108e-14 EdH/spin 06-20 kosta M61^16+1
...

[/code]
from the best poly score

Gimarel 2021-01-12 17:58

My two best polys:[CODE]# norm 1.742952e-18 alpha -7.250051 e 2.198e-14 rroots 3
skew: 28373488.56
c0: -36289735353276157276527068902255490733853568
c1: 4334673014099287941043579868135706812
c2: -407368329922469859774625933879
c3: 1748229736159966635685
c4: 1727908216745850
c5: 3603600
Y0: -6922993593233741704051927826032280373
Y1: 22185579502781921755367
# norm 1.726629e-18 alpha -6.853313 e 2.170e-14 rroots 3
skew: 20914898.48
c0: -9021236659014956031207533880822959360049758
c1: 3167077071049478779652836766937560993
c2: -376983337640966863439394185914
c3: 12016958982774661681885
c4: 1754473649639850
c5: 3603600
Y0: -6922993560523700440001809682589016812
Y1: 22185579502781921755367[/CODE]

VBCurtis 2021-01-12 19:06

My poly search will complete Wed afternoon (USA-westcoast time).

Without Gimarel, I'd be pretty convinced msieve was no longer competitive for poly select. Nice work, again!

I'll do a bit of test-sieving, in particular tinkering with adjust-strategy 2 that henryzz suggested a month ago and Charybdis has found speeds many jobs but not all jobs.

I ought to have a CADO server running Thursday evening for this job; I'll post the port number here when it's ready. This job with be I=15, default number of threads 4 per client (you can override if you wish). I'll be aiming about 30 cores at the job personally.

EdH 2021-01-12 19:47

[QUOTE=charybdis;569082]. . . I tried spinning with EdH's script but didn't get an improvement.[/QUOTE]Just to be fair to Max0526. It is his spin. I (and swellman) just got with him to work on bringing it to a wider availability and try to automate it for ease of use. Also, there is more to the process that hasn't been brought forth yet. So, there may still be some room for improvement.

Gimarel 2021-01-12 20:10

[QUOTE=VBCurtis;569097]Without Gimarel, I'd be pretty convinced msieve was no longer competitive for poly select. Nice work, again![/QUOTE]
I use the sizeoptimisation of cado, so it's not entirely msieve.
And I modified a few parameters of msieve.

VBCurtis 2021-01-14 00:10

My poly score was about 5% worse than Charybdis'.

Charybdis has test-sieved his best vs both of Gimarel's, and declares the results (in speed and yield both) a dead heat between his best and Gimarel's higher-scoring poly. I'll do some testing this evening, pick one of the polys, and will get the CADO server running Thursday some time.

Since I am not a security expert and don't know how to harden my machine well, I prefer to leave the server address not on this public forum. Please PM me if you wish to contribute to this factorization, and I'll send you the server address and port number once I have it up and running (should be same as it was for the GNFS-207 team job).

VBCurtis 2021-01-14 07:23

The machine is up and serving workunits. PM me for server info.

Please remember to include the flag --bindir=./build/{yourmachinename}, as the las binary cado-nfs-client downloads from my machine may crash on your machine's architecture.

Also, recall --override t {n} changes the default 4 threads to however many threads you'd like your client to run.

Q-range is 2000 per workunit, so they run pretty fast; you won't waste much work if you stop the clients (say, Ed's machines when they nap for the night).

Initial ETA on the machine is around 90 days, with 20 cores running. That's at Q=30M.

EdH 2021-01-14 13:41

[QUOTE=VBCurtis;569249]. . . (say, Ed's machines when they nap for the night). . .[/QUOTE]Just to note that my slumbering machines are supposed to go dormant after finishing a WU. If you see (m)any crashed WUs for me, please let me know.

EdH 2021-01-14 14:30

I've turned loose a bunch of my machines, but I forgot to check on memory needs. I may have to remove a few. I've forgotten how they fail. I will try to look for faults, but if you see some that should be dropped, let me know.

charybdis 2021-01-14 14:50

My machines are up and running.

[QUOTE=EdH;569272]I've turned loose a bunch of my machines, but I forgot to check on memory needs. I may have to remove a few. I've forgotten how they fail. I will try to look for faults, but if you see some that should be dropped, let me know.[/QUOTE]

Adjust-strategy 2 does use a bit more memory than the default, but each process should still fit within 4GB.

EdH 2021-01-14 15:59

[QUOTE=charybdis;569274]. . .
Adjust-strategy 2 does use a bit more memory than the default, but each process should still fit within 4GB.[/QUOTE]Thanks! All of the machines I connected, but one, appeared to be running OK. My strategy for simplicity of a single script, uses all available threads in a single process for all but one of my machines. That machine has 24 threads and lots of memory, so I treat it differently.

VBCurtis 2021-01-14 17:37

'top' shows 2.6GB per process on my machine. Glad these params work on all your machines, too!

We've already found 20M relations, and ETA is Jan 25th. Naturally, that ETA will stretch out as we get to higher Q, and Ed's machines don't run 24 hrs a day.

I set rels-wanted to 540M; I think with the tight bounds on mfb0 and lambda0 that's plenty to get a matrix in the 30M range.

VBCurtis 2021-01-15 00:39

56M relations, ETA is now late Wednesday. Surely that will lengthen when Ed's farm naps, but that's impressive for a day's work!

Q is approaching 47M, so yield is 3.3 so far.

Pretty cool to see what we can do in a week, even without nfs@home!

RichD 2021-01-15 01:32

I added a couple Sandy Bridge Linux systems as "localhost" today.

Max0526 2021-01-15 15:46

C191 poly
 
[QUOTE=charybdis;569082]The best from my run, by some margin, is
[code]n: 28653448550423582065969107478752150888235561427247664110066527668159681894430155521476413904809897034030256344677647748498114768327844226758360594247371333432148627535343266861185808662829021
skew: 61160593.83
c0: -782066809369899992846921498396677009099575965
c1: 129978232145697650128010452670728125268
c2: 1136403038232649046041648047998
c3: -50366553898678575377716
c4: 291334628577375
c5: 2637600
Y0: -6410674956398678515938860432260175622
Y1: 195016503229310813759987119[/code]
Cownoise scores this as 2.023e-14. I tried spinning with EdH's script but didn't get an improvement.[/QUOTE]

Try EdH's script on a scaled poly: :stirpot: [code]R0: -6411935581103770910754613672194879827
R1: 390033006458621627519974238
A0: -97543900524578359780002030035925410223410400
A1: 13585100386688885228817525991732114606
A2: 544764545832419180052744447527
A3: -28398693293522783165108
A4: 206084824917375
A5: 5275200
skew: 40462841.77347
# size 1.581e-18, alpha -7.281, combined = 2.372e-14 rroots = 3[/code]
Maybe the script will improve it even more.

swellman 2021-01-15 16:07

Too late for use in this GNFS job, but an impressive new high score for a c191. And it could improve even further with the script if someone wants run it.

VBCurtis 2021-01-15 16:44

Max's magic beats the CADO group's score for RSA-190, on a C191.
Nice find!

Max0526 2021-01-15 17:33

[QUOTE=VBCurtis;569379]Max's magic beats the CADO group's score for RSA-190, on a C191.
Nice find![/QUOTE]
It only took us three years to shave a digit off then! :-)

henryzz 2021-01-16 00:17

[QUOTE=EdH;569272]I've turned loose a bunch of my machines, but I forgot to check on memory needs. I may have to remove a few. I've forgotten how they fail. I will try to look for faults, but if you see some that should be dropped, let me know.[/QUOTE]

You could fiddle with some values for -bkthresh1. I have seen this reduce memory usage a lot in large memory cases. Reducing the memory too much can slow it down.

EdH 2021-01-16 02:50

[QUOTE=henryzz;569415]You could fiddle with some values for -bkthresh1. I have seen this reduce memory usage a lot in large memory cases. Reducing the memory too much can slow it down.[/QUOTE]Thanks, but as far as I can tell, all the machines I'm connecting are working fine this time. I will try to keep your note in mind for the future.

EdH 2021-01-16 02:54

[QUOTE=Max0526;569366]Try EdH's script on a scaled poly: :stirpot: [code]R0: -6411935581103770910754613672194879827
R1: 390033006458621627519974238
A0: -97543900524578359780002030035925410223410400
A1: 13585100386688885228817525991732114606
A2: 544764545832419180052744447527
A3: -28398693293522783165108
A4: 206084824917375
A5: 5275200
skew: 40462841.77347
# size 1.581e-18, alpha -7.281, combined = 2.372e-14 rroots = 3[/code]Maybe the script will improve it even more.[/QUOTE]It came up with some different coefficients, but no improvement for score:
[code]
Y0: -6410674956983728025626792873540136979
Y1: 390033006458621627519974238
c0: 48879199956536637644447993225484888790155680
c1: -16247278165909757604391180995060813806
c2: -284100872882912466554514368423
c3: 25183278697346940461108
c4: -291334589013375
c5: -5275200
skew: 30580296.915
# lognorm 57.86, E 50.58, alpha -7.28 (proj -1.72), 3 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16)=2.363e-14
[/code]cownoise: 40406010.90765 2.37151357e-14

VBCurtis 2021-01-16 17:48

Status update:
Q=30-114M complete. 237M relations collected, out of a target of 540M. 44% complete.
ETA is Tuesday midday. Estimated ending Q is 240M.

VBCurtis 2021-01-18 20:06

Q=30-205M complete. 429M raw relations so far, a hair under 80% complete.
ETA is Tuesday 4:30pm USA-pacific time.
Estimated ending Q is 255-260M; current yield is close to 2.0.

charybdis 2021-01-18 21:23

Just thrown in 45 more cores to help speed things along to the finish line.

VBCurtis 2021-01-20 00:44

We reached 540M relations, and CADO crashed on the first filtering step with this error:
[code]Error:Filtering - Duplicate Removal, removal pass: terminate called after throwing an instance of 'renumber_t::corrupted_table'
Error:Filtering - Duplicate Removal, removal pass: what(): Renumber table is corrupt: cannot find data for prime 0x28e17329 ; note: isprime(p)==0
Error:Filtering - Duplicate Removal, removal pass: Note: above 2^32 ideals or relations, add FLAGS_SIZE="-DSIZEOF_P_R_VALUES=8 -DSIZEOF_INDEX=8" to local.sh[/code]

I've seen this error on small (~C110) jobs with recent checkouts on another machine, which caused me to stick with the May-2020 version that works on that other machine. I used a fresh git-checkout for this job, and am disappointed to see that it's a bigger bug than I guessed.

In our case, it shouldn't much matter- I'm going to use msieve for postprocessing anyway. I'll let y'all know what I've got after filtering.

Edit: remdups leaves me 368.0M uniques, a pleasant 68%. Q-range was 30M to 264.6M; yield fell off quite a bit at the end. If I did this size of number again, I think I'd use A=30 for Q=30-50M to boost yield at the bottom, and I'd increase lim's by 25% or so (I used 130/180M on this job).

VBCurtis 2021-01-20 04:19

First filtering pass:

[code]Tue Jan 19 20:02:16 2021 begin with 368134973 relations and 408812974 unique ideals
Tue Jan 19 20:08:56 2021 reduce to 136356813 relations and 139341382 ideals in 25 passes[/code]

That leaves us a bit short of the necessary relations. I've restarted the CADO server, asking for 570M total relations (30M more).

Looks like Charybdis' machines clients were still waiting, as they got work immediately. Mine are running too, so if you've moved on to other tasks don't worry about your clients. 30M relations won't take long.

charybdis 2021-01-20 13:12

[QUOTE=VBCurtis;569706]First filtering pass:

[code]Tue Jan 19 20:02:16 2021 begin with 368134973 relations and 408812974 unique ideals
Tue Jan 19 20:08:56 2021 reduce to 136356813 relations and 139341382 ideals in 25 passes[/code]

That leaves us a bit short of the necessary relations. I've restarted the CADO server, asking for 570M total relations (30M more).

Looks like Charybdis' machines clients were still waiting, as they got work immediately. Mine are running too, so if you've moved on to other tasks don't worry about your clients. 30M relations won't take long.[/QUOTE]

That looks a lot more than 30M relations short to me. To get a good matrix I'd want the initial number of relations to be a bit bigger than the number of ideals, so you probably want 100M more raw relations to get another ~70M unique.

Would be a good idea to set rels_wanted to something like 800M so the clients keep going, and keep running msieve filtering until you get a decent matrix.

PS. most of my machines rebooted overnight to install updates, apologies

EdH 2021-01-20 14:30

I'll be tasking my farm back to you shortly. The "night owl" machines were running a night project.

VBCurtis 2021-01-20 17:24

OK, I changed target rels to 670M. I'll try msieve-filtering this afternoon at ~570M raw rels and report back.

VBCurtis 2021-01-21 04:10

578M raw relations got me 391M uniques, and a 39.98M matrix at TD=84.

There was a deficit of 27M in relations vs ideals in the first step of filtering, so I'm continuing to sieve. I think 610M or so should be enough, and we're at 600M now, so feel free to repurpose your machines any time.

I'll leave the server to run till morning (USA-pacific time), and will run a final filtering job with however many relations I have when I wake up.

VBCurtis 2021-01-21 15:37

597M relations built a 34.4M matrix at TD 100 overnight.

I lost internet connection to the CADO server overnight, so we quit serving workunits at 615M relations. That'll be enough, so we're "officially" done sieving. I'll report on the actual matrix size and ETA later today.

Edit: 615M raw relations yielded 414M uniques, enough for a 31M matrix at TD 116. No ETA yet, I'll edit again in a few hours once the matrix is running.

EDIT2: ETA is just under 10 days, running on all 20 cores without MPI. I recently updated that machine to Ubuntu 20.04, so I may try to get the MPI fix applied and see how msieve-MPI performs on it.

EdH 2021-01-22 18:31

[QUOTE=VBCurtis;569777]. . . I recently updated that machine to Ubuntu 20.04, so I may try to get the MPI fix applied and see how msieve-MPI performs on it.[/QUOTE]You shouldn't need to fix anything (if you're referring to openmpi), but you will probably need to have the same Ubuntu version on any other machines you plan to include in your cluster. The following should be the only thing necessary and it should really already be accomplished if you already had openmpi installed in the earlier version:
[code]
apt update
apt install openmpi-bin openmpi-common libopenmpi-dev
[/code]

VBCurtis 2021-01-22 19:57

No plans to use another machine- simply to MPI across the two sockets, as the reduction in cross-messaging is reported to give 15-20% speedup compared to the current "msieve -t 20" for the two ten-core sockets.

I'll play with it tonight- gotta go pick up a Ferrari this afternoon! The very best kind of errand, especially when it's not my car. :)

swellman 2021-02-06 21:36

Yoyo@Home is finishing the batch of base-2 Cunningham numbers (1987 edition). Should be completely finished in another week or two. Each composite in this group has survived 12000 curves @B1 = 850e6. One iteration down, five more to go. Of course that’s not counting any earlier ECM work.

Next step? Following a strategy of finish-what-you-started, I can just queue all the surviving composites back into Yoyo for another round of 0.16t65 (ETA ~13 months). Or we could select a small subset of “interesting” cases for focused ECM and eventual NFS by Greg on NFS@Home. There are two quartics - start with them? Or others?

With the recent extension of the base-2 tables, do we leave some of Yoyo’s resources available to work that effort? Or is ECM on Yoyo even needed with this latest batch of Cunninghams?

My opinion is to split the baby - focus on a few of the 1987 base-2 Cunninghams and leave room for other work. Feeds Greg a few more candidates by this fall and avoids flooding Yoyo’s queue. But I’m certainly no Solomon...

Batalov 2021-02-06 22:00

1 Attachment(s)
I would suggest the low-hanging fruit of base-2 extensions. Those might yield some hits because SSW himself just pulled a p56, so there are some p5x factors [URL="https://homes.cerias.purdue.edu/~ssw/cun/xtend/appc7.txt"]in this set[/URL] sans 3+ that are already done : [url]https://homes.cerias.purdue.edu/~ssw/cun/xtend/up7.txt[/url]

But there is a good argument to continue hitting hard on Base-2 main, too. Three inch punch! It really worked for her in the end!

charybdis 2021-02-19 01:42

[QUOTE=VBCurtis;569777]
EDIT2: ETA is just under 10 days, running on all 20 cores without MPI. I recently updated that machine to Ubuntu 20.04, so I may try to get the MPI fix applied and see how msieve-MPI performs on it.[/QUOTE]

It's been a few weeks. Guessing you put this on hold to work on something else?

VBCurtis 2021-02-19 02:19

1 Attachment(s)
Whoops! Sorry, I was so eager to jump on the quartic from the Cunningham extensions that I forgot to post factors when the matrix finished:
[code]Sun Jan 31 00:34:01 2021 p82 factor: 8839572613855891331570438385350118192600962609702949549624218311059899144936218181
Sun Jan 31 00:34:01 2021 p109 factor: 3241497049926344989867894223963855082968056873457044230450152346084471878024227245959456141481343243331475641[/code]

Log attached.

charybdis 2021-02-19 02:26

Good job everyone! 20 days from start to finish isn't too shabby :smile:

Curtis, have you sent the factors to Sam Wagstaff?

VBCurtis 2021-02-19 02:56

Negative. First report anywhere is two posts up. :no:

charybdis 2021-02-19 03:02

OK, might as well send them myself then...


All times are UTC. The time now is 23:19.

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