Polynomial Request Thread - Page 105

VBCurtis · 2018-04-06, 21:18

I sort-of-tied Rich's best msieve poly:

Code:

# norm 2.736995e-19 alpha -8.997274 e 6.700e-15 rroots 5
skew: 746548923.72
c0: -75731743411081676572988812351153474024302171082752
c1: 306455460524824208896007078950091210537232
c2: 2098111595374077238498473747669972
c3: -5639730128380960746645544
c4: -4689342804907975
c5: 601692
Y0: -128280183283170652823577025991627503511
Y1: 18854473277774059343

Still going through 500k-1M on both CADO and msieve.

swellman · 2018-04-06, 21:26

I’ve been playing with parameters to maximize efficiency using msieve. I’ll use them to punch through to 0.5M.

Max0526 · 2018-04-06, 21:50

Quote:

Originally Posted by VBCurtis

I sort-of-tied Rich's best msieve poly:

Code:

# norm 2.736995e-19 alpha -8.997274 e 6.700e-15 rroots 5
skew: 746548923.72
c0: -75731743411081676572988812351153474024302171082752
c1: 306455460524824208896007078950091210537232
c2: 2098111595374077238498473747669972
c3: -5639730128380960746645544
c4: -4689342804907975
c5: 601692
Y0: -128280183283170652823577025991627503511
Y1: 18854473277774059343

Still going through 500k-1M on both CADO and msieve.

And this one spins into our test-sieve candidate number three:

Code:

Y0: -128280183283175038524897222460023160055
Y1: 37708946555548118686
c0: -4737682113152575563349555619330383737950004367008
c1: 38184808779885988297067997820651742275074
c2: 525511402998103765746708260854197
c3: -2817683344110966477947732
c4: -4690042596771655
c5: 1203384
skew: 374011545.23
# size 2.697e-019, alpha -8.824, combined = 7.570e-015 rroots = 5

swellman · 2018-04-08, 10:48

Quote:

Originally Posted by Max0526

I think it would be beneficial (for my scaling afterwards, anyways) to have msieve's c5 coefficient to be a multiple of 60 (same as in CADO) instead of the default multiple of 12.
Is it possible to modify?

Quote:

Originally Posted by chris2be8

You could write a script to run:
msieve -np1 60,60 ...
msieve -np1 120,120 ...
msieve -np1 180,180 ...

over whatever range you wanted.

Chris

Is this worth pursuing? Have all recent high scoring polys found using msieve-GPU/CADO been a multiple of 60?

I could modify a part of my remaining run using a script as Chris suggests. No guarantees of a worthwhile hit but it cuts the search space by 80%.

As to my parameters, I ran a bit of iteration trying to balance speed and yield (i.e. polys/hr). Still trying to intuitively get it, but I did find a set of parameters that seemed to work best. Unfortunately nothing will make my old laptop GPU suddenly start producing like a latest gen rig. So I’m grinding through c5 < 0.5M now but it will take a few more weeks to finish.

Gimarel · 2018-04-08, 13:14

If you compile msieve yourself, you can change the multiplier of the leading coefficient in

Code:

gnfs/poly/stage1/stage1.h

Change the value of HIGH_COEFF_MULTIPLIER to

Code:

#define HIGH_COEFF_MULTIPLIER 60

Max0526 · 2018-04-08, 14:03

My best poly so far. It doesn't seem to scale, too bad.

Code:

Y0: -195365504676471523605765805321679260453
Y1: 646375052504307000329
c0: -92860587798543236913935845305022995692578935202560
c1: 83258718746728188594033447624453329948388
c2: 1366184475883317396422140549591417
c3: -778530084534309572957233
c4: 964584764976558
c5: 146880
skew: 1223281864.82
# size 2.549e-019, alpha -9.244, combined = 7.317e-015 rroots = 3

swellman · 2018-04-08, 14:51

Code:

R0: -182324558401028859022313707410656977574
R1: 2088734617065907253
A0: 9407119911616431138021081681550586197240792700551
A1: 11351997243696540225891008583027166486936
A2: -504061319264223156461502891116793
A3: -444480432486007993252736
A4: 2497913106433102
A5: 103740
skew 512382524.73, size 1.925e-19, alpha -7.522, combined = 6.073e-15 rroots = 5

Embarrassingly low score but my best to date.

Max0526 · 2018-04-08, 16:25

Quote:

Originally Posted by swellman

Code:

R0: -182324558401028859022313707410656977574
R1: 2088734617065907253
A0: 9407119911616431138021081681550586197240792700551
A1: 11351997243696540225891008583027166486936
A2: -504061319264223156461502891116793
A3: -444480432486007993252736
A4: 2497913106433102
A5: 103740
skew 512382524.73, size 1.925e-19, alpha -7.522, combined = 6.073e-15 rroots = 5

Embarrassingly low score but my best to date.

This one doesn't scale either. :-(

swellman · 2018-04-08, 23:41

The C205 in aliquot sequence 276 looks ready for poly select. (Unless I’ve missed another candidate).

I still have a few weeks to complete my part on the C197, though others may be ready to move on sooner.

ETA: Forgot that Max has a C200 that needs solving.

VBCurtis · 2018-04-09, 06:15

I prefer aliquot sequences in general, so I'd like to do the C205 next. I'm open to being convinced otherwise, e.g. Max has private resources lined up to do the sieving and wants a poly by 1 May to feed them.

I'll be moving on from 4788-C197 by Friday.

Max0526 · 2018-04-09, 17:47

Quote:

Originally Posted by VBCurtis

I prefer aliquot sequences in general, so I'd like to do the C205 next. I'm open to being convinced otherwise, e.g. Max has private resources lined up to do the sieving and wants a poly by 1 May to feed them.

I'll be moving on from 4788-C197 by Friday.

No C200 rush for now. I'm up for C205 polyselect any time.

2018-04-08, 13:14	#1149
Gimarel Apr 2010 A7₁₆ Posts	If you compile msieve yourself, you can change the multiplier of the leading coefficient in Code: gnfs/poly/stage1/stage1.h Change the value of HIGH_COEFF_MULTIPLIER to Code: #define HIGH_COEFF_MULTIPLIER 60

2018-04-08, 23:41	#1153
swellman Jun 2012 2²·773 Posts	C205 from AS276_i2122 The C205 in aliquot sequence 276 looks ready for poly select. (Unless I’ve missed another candidate). I still have a few weeks to complete my part on the C197, though others may be ready to move on sooner. ETA: Forgot that Max has a C200 that needs solving. Last fiddled with by swellman on 2018-04-09 at 00:02

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
GIMPS wiki account request thread	ixfd64	mersennewiki	169	2018-09-21 05:43
Polynomial Discriminant is n^k for an n-1 degree polynomial	carpetpool	Miscellaneous Math	14	2017-02-18 19:46
Lost Prime Raider password request thread	cheesehead	Forum Feedback	6	2009-07-28 13:02
Polynomial	R.D. Silverman	NFSNET Discussion	13	2005-09-16 20:07
Deutscher Thread (german thread)	TauCeti	NFSNET Discussion	0	2003-12-11 22:12

2018-04-06, 21:26	#1146
swellman Jun 2012 2²·773 Posts	I’ve been playing with parameters to maximize efficiency using msieve. I’ll use them to punch through to 0.5M.

2018-04-09, 06:15	#1154
VBCurtis "Curtis" Feb 2005 Riverside, CA 2×2,437 Posts	I prefer aliquot sequences in general, so I'd like to do the C205 next. I'm open to being convinced otherwise, e.g. Max has private resources lined up to do the sieving and wants a poly by 1 May to feed them. I'll be moving on from 4788-C197 by Friday.