Polynomial Request Thread - Page 101

VBCurtis · 2018-03-28, 04:17

Thanks Max! I'm testing the 1.226 now for best parameters. I'll then use the same parameters to test the other two, will let you know what I chose.

This round to CADO by knockout; if other C180+ poly searches are this lopsided, I'll be retiring my GPU-msieve efforts.

Which are we on to next? The Aliq-276 C205? There are a couple C190's in various threads; I'm presently quite excited to try CADO on a 190+!

Max0526 · 2018-03-28, 05:31

Fun fact: our recent C179 poly is better than the old record C178 poly!
CADO! CADO!

swellman · 2018-03-28, 18:24

I just finished a msieve run from 1.5M-2M, best was 9.77. Blah. Currently running 2-3M with hope —> 0.

@VBCurtis, if you want to call it for CADO 1.2+, I will gladly pull the plug on my current sweep. At this point I would have to generate a 1.3+ to justify changing horses. If you want me to run it out, it should be done by Monday night.

VBCurtis · 2018-03-28, 20:05

No need to continue, Sean. A trio of 1.2's is plenty for this number- as Max noted, these are records for C178, let alone C179!

Thanks everyone for the assistance- I'll report params at the end of the week. When testing 15f, 34LP is substantially faster than 33LP (like sec/rel is around 52-55%). So, if I get factmsieve.py altered to work with 15f by not changing lim's for each .job file, I may try my first 34LP factorization! lim's will likely be 134M, sieve range 20M-75M. 33LP yield is around 9, while 34LP yield is 17-18. Was planning for 600M raw 33LP relations, now figuring 1000M-1050M 34LP raw rels.

swellman · 2018-03-28, 21:45

Sounds good - stopping msieve immediately. Looking forward to observing your sieving effort, I’ll help where I can. Not sure my 32Gb machine can handle a LP 34 job in LA.

Ready to help on the next poly search once the composite is specified.

For the record, I’ve got a GNFS C178 for our queue sometime in the future.

Max0526 · 2018-03-28, 22:04

Quote:

Originally Posted by VBCurtis

Which are we on to next? The Aliq-276 C205? There are a couple C190's in various threads; I'm presently quite excited to try CADO on a 190+!

My C200 will wait until C190's are selected. I guess the Aliq-276 C205 will wait too. Please pick one of the C190's for us, whichever is more urgent.

swellman · 2018-03-29, 00:12

I care not which sequence we choose, but here’s a summary of the four public aliquot sequences:

AS 3408 - C192 still needs a t60 run, but potentially a long slog ahead, sequence has been sidelined for the foreseeable future by group consensus

AS 3366 - C195 but it’s already being sieved by NFS@Home, Tom Womack to postprocess

AS 276 - C205 still working through(?) a partial t70. (Unclear if yoyo has figured a way to handle t70 Stage 2).

AS 4788 - C197 with ECM to partial t65 (goal) complete. Ready for poly.

Max0526 · 2018-03-29, 00:29

Please count my vote for AS 4788 - C197. It's time to override that 7.240e-15. Luckily we are smarter now than in September 2016. I suggest posting a poly with E>=7e-15 as our first baseline.
8-handle ho!

VBCurtis · 2018-03-29, 01:19

Agreed, 4788 C197 next.
Max, please let me know which ad-range you run for CADO; I again plan to use nq=15625, and believe that's more useful on lower ad-ranges (you're welcome to do so!). I've been bumping default P up by 15-20% (in round numbers, e.g. 5M -> 6M), dunno if that's helpful.

For msieve, I'll start at 5M; I have some GPU-ECM going, but the GPU will be free by Friday.

swellman · 2018-03-29, 01:32

I’ll run up to 2M using msieve GPU on C197 of AS 4788, starting tonight.

RichD · 2018-03-30, 00:06

I'll start a run of 4-5M on C197 from AS 4788 using msieve-GPU.

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

2018-03-28, 04:17	#1101
VBCurtis "Curtis" Feb 2005 Riverside, CA 2×2,437 Posts	Thanks Max! I'm testing the 1.226 now for best parameters. I'll then use the same parameters to test the other two, will let you know what I chose. This round to CADO by knockout; if other C180+ poly searches are this lopsided, I'll be retiring my GPU-msieve efforts. Which are we on to next? The Aliq-276 C205? There are a couple C190's in various threads; I'm presently quite excited to try CADO on a 190+!

2018-03-28, 05:31	#1102
Max0526 "Max" Jun 2016 Toronto 19×47 Posts	Fun fact: our recent C179 poly is better than the old record C178 poly! CADO! CADO!

2018-03-28, 18:24	#1103
swellman Jun 2012 11·281 Posts	I just finished a msieve run from 1.5M-2M, best was 9.77. Blah. Currently running 2-3M with hope —> 0. @VBCurtis, if you want to call it for CADO 1.2+, I will gladly pull the plug on my current sweep. At this point I would have to generate a 1.3+ to justify changing horses. If you want me to run it out, it should be done by Monday night.

2018-03-28, 20:05	#1104
VBCurtis "Curtis" Feb 2005 Riverside, CA 2×2,437 Posts	No need to continue, Sean. A trio of 1.2's is plenty for this number- as Max noted, these are records for C178, let alone C179! Thanks everyone for the assistance- I'll report params at the end of the week. When testing 15f, 34LP is substantially faster than 33LP (like sec/rel is around 52-55%). So, if I get factmsieve.py altered to work with 15f by not changing lim's for each .job file, I may try my first 34LP factorization! lim's will likely be 134M, sieve range 20M-75M. 33LP yield is around 9, while 34LP yield is 17-18. Was planning for 600M raw 33LP relations, now figuring 1000M-1050M 34LP raw rels.

2018-03-28, 21:45	#1105
swellman Jun 2012 11·281 Posts	Sounds good - stopping msieve immediately. Looking forward to observing your sieving effort, I’ll help where I can. Not sure my 32Gb machine can handle a LP 34 job in LA. Ready to help on the next poly search once the composite is specified. For the record, I’ve got a GNFS C178 for our queue sometime in the future.

2018-03-29, 00:12	#1107
swellman Jun 2012 11·281 Posts	I care not which sequence we choose, but here’s a summary of the four public aliquot sequences: AS 3408 - C192 still needs a t60 run, but potentially a long slog ahead, sequence has been sidelined for the foreseeable future by group consensus AS 3366 - C195 but it’s already being sieved by NFS@Home, Tom Womack to postprocess AS 276 - C205 still working through(?) a partial t70. (Unclear if yoyo has figured a way to handle t70 Stage 2). AS 4788 - C197 with ECM to partial t65 (goal) complete. Ready for poly.

2018-03-29, 00:29	#1108
Max0526 "Max" Jun 2016 Toronto 19×47 Posts	Please count my vote for AS 4788 - C197. It's time to override that 7.240e-15. Luckily we are smarter now than in September 2016. I suggest posting a poly with E>=7e-15 as our first baseline. 8-handle ho!

2018-03-29, 01:19	#1109
VBCurtis "Curtis" Feb 2005 Riverside, CA 1001100001010₂ Posts	Agreed, 4788 C197 next. Max, please let me know which ad-range you run for CADO; I again plan to use nq=15625, and believe that's more useful on lower ad-ranges (you're welcome to do so!). I've been bumping default P up by 15-20% (in round numbers, e.g. 5M -> 6M), dunno if that's helpful. For msieve, I'll start at 5M; I have some GPU-ECM going, but the GPU will be free by Friday.

2018-03-29, 01:32	#1110
swellman Jun 2012 3091₁₀ Posts	I’ll run up to 2M using msieve GPU on C197 of AS 4788, starting tonight.

2018-03-30, 00:06	#1111
RichD Sep 2008 Kansas D3F₁₆ Posts	I'll start a run of 4-5M on C197 from AS 4788 using msieve-GPU.