mersenneforum.org 3,748+ c204 Smaller-but-Needed
 Register FAQ Search Today's Posts Mark Forums Read

2021-09-08, 00:29   #56
swellman

Jun 2012

3,203 Posts

Quote:
 Originally Posted by charybdis I used some top secret methods to figure out the pre-sizeopt leading coefficients of Ed's polys, and he appears to have searched some range from 100M upwards - perhaps to 110M? - at incr=420 and a very large P value of 16M.
No, I believe you’ll find that EdH found that particular poly by guessing and he got really lucky.

Fair enough, I was only counting the declared ranges. You’re right of course EdH must have searched high up the mountain.

Quote:
 I'm not convinced we can't do better here. 2.6 feels a bit low, and the sample size of c204s is too small for the record to be particularly meaningful. The switch to incr=4620 was made quite early, so I'm going to break one of the fundamental rules of poly selection and duplicate some existing work by doing 500k-5M at P=8M, incr=420. I may go further if this looks promising.
No argument from me - we likely can do better. I think EdH’s poly was decent and would work quite well, but there’s certainly room to improve without needing a monster machine.

Personally, I’ve gotten in the habit of jumping to higher admin, say 30M+, and using incr=4620. Often generates reasonable polys within a week or two. But your recent examples have made a believer out me - stick with 420 for as long as you can! I’m using it locally on a G193 job now. Slow but sure can win the race, or at least generate better quality polynomials.

2021-09-09, 22:51   #57
charybdis

Apr 2020

2×251 Posts

Quote:
 Originally Posted by charybdis I'm going to break one of the fundamental rules of poly selection and duplicate some existing work by doing 500k-5M at P=8M, incr=420.
This didn't get anywhere, but I also ran 500k-10M at P=5M, incr=420, which gave this:

Code:
n: 533439167600904850230361756102700151678687933392166847323827307497363839257031077774321424872955045754669625577486179222154434651598903112919949771321416511589029559325246084363632977829645558547714072241
skew: 338012013.442
c0: 39995364790245999480244642774391120587683667437750
c1: 663343854938705696739096400303369099559943
c2: -2554996497228376927871276092481512
c3: -16634883235399116321901043
c4: 14147480163732582
c5: 2484720
Y0: -3361764766011432381851898081097482786808
Y1: 1134911427375579461125271
Score is 2.658e-15 so a slight improvement over Ed's poly, but close enough that test-sieving would be needed. I'll keep going with 10M-20M.

 2021-09-12, 15:30 #58 charybdis     Apr 2020 7668 Posts Continued up to 35M. Best score was 2.807e-15, probably good enough for sieving? Code: n: 533439167600904850230361756102700151678687933392166847323827307497363839257031077774321424872955045754669625577486179222154434651598903112919949771321416511589029559325246084363632977829645558547714072241 skew: 190550095.962 c0: -4270861597978000348273667478612159513444633369624 c1: 301982688153622958148557487129513846465195 c2: 718476090251496284529736703791030 c3: -12811377464083267088449895 c4: -31157665354464146 c5: 55931400 Y0: -1803400746044964489961544593826293030270 Y1: 15050066776523075044713587
 2021-09-12, 18:55 #59 kruoli     "Oliver" Sep 2017 Porta Westfalica, DE 11×61 Posts Are those two last polynomials already spun?
 2021-09-12, 20:44 #60 charybdis     Apr 2020 2×251 Posts I haven't run anything beyond basic CADO poly selection. Feel free to try spinning with EdH's script. I think Max does some more things that are known only to him.
 2021-09-12, 21:04 #61 kruoli     "Oliver" Sep 2017 Porta Westfalica, DE 11·61 Posts In this case I'd like to ask Max if he would be willing to apply his special sauce here?
2021-09-12, 21:12   #62
EdH

"Ed Hall"
Dec 2009

2×2,011 Posts

Quote:
 Originally Posted by charybdis I haven't run anything beyond basic CADO poly selection. Feel free to try spinning with EdH's script. I think Max does some more things that are known only to him.
I've given up on my script, at least for CADO-NFS polys. Max and Gimarel do stuff with Msieve that I haven't figured out yet. If I ever do, I'll add it to the script. Also, Max does some re-balancing of coefficients with Maple that I haven't been able to program.

 2021-09-12, 22:05 #63 VBCurtis     "Curtis" Feb 2005 Riverside, CA 138916 Posts Judging by the records for 200-201-202, a 3.0 should be possible. Like Kruoli, I'll hope for some lucky spin!
2021-09-13, 22:02   #64
Max0526

"Max"
Jun 2016
Toronto

38A16 Posts

Quote:
 Originally Posted by kruoli In this case I'd like to ask Max if he would be willing to apply his special sauce here?
Got a PM from swellman. On it!

 2021-09-14, 14:24 #65 Gimarel   Apr 2010 18410 Posts I'll try to get a better poly with msieve-gpu.
 2021-09-14, 18:14 #66 Plutie   "Evan" Dec 2020 Montreal 4716 Posts I've been trying to run some polyselect with the individual sopt/ropt binaries. Is there a good way to sort out a certain amount of the highest scoring polynomials after the size optimization phase? I found a script from Dubslow from a few years ago, but it doesn't seem to work on the files I have. I have a feeling it's something to do with the fact that CADO doesn't output Murphy-E values from the sopt run, only the ropt run.

 Similar Threads Thread Thread Starter Forum Replies Last Post fivemack Factoring 3 2017-09-19 08:52 skan YAFU 6 2013-02-26 13:57 akruppa Factoring 114 2012-08-20 14:01 fortega Data 2 2005-06-16 22:48 marc Factoring 6 2004-10-09 14:17

All times are UTC. The time now is 13:15.

Mon Oct 25 13:15:04 UTC 2021 up 94 days, 7:44, 0 users, load averages: 1.80, 1.30, 1.18