GNFS targets which need more ECM

[swellman]

Stage 2 of C174_136_69 completed for all 18,000 curves with no factors found.

[swellman]

Originally Posted by fivemack

I have a shiny new GTX1080 on the table downstairs, which ought to be quite good at GPU-ECM. May I reserve C200_139_113 ?

My ecm-toy tool suggests (on the assumption that it's already had 7610@43e6) a recipe of 5700@43e6 then 1400@110e6 then 34200@260e6 then 2500@850e6 - though my tool does not know about GPUs so may have a different recipe once I've done some timings on the GTX1080.

Not sure if Andrey saw this, but bumping it in case fivemack still wants to work on C200_139_113.

It did previously survive 7600 curves @B1=43e6.

[swellman]

Originally Posted by swellman

Here is the remainder of my list. Not sure which (if any) are practical GNFS targets.

C201_137_134
C202_147_116
C203_145_119
C203_146_107
C203_147_104
C203_137_127
C203_142_87
C204_147_118
C206_139_123
C207_143_127
C208_145_99
C211_137_135
C213_147_128
C213_141_113
C214_143_135
C214_143_119
C218_142_133
C222_149_141
C228_145_141
C230_149_136

Where should we draw the line on GNFS? I believe the 16e siever of NFS@Home has handled upwards of 220 digits but that's a huge undertaking that Greg may not want to entertain.

Is 208 inclusive a feasible value? I really have no idea.

[swellman]

Ryan is reserving C182_125_121.

[swellman]

Originally Posted by swellman

Ryan is reserving C182_125_121.

Factored by ECM. p59*p123

Input number is
19638514892326880715049637947667358096563021648596014807027282348855860502922717331324961311298405146927437844559477346581035562867109341355691061338230447718948577630638933884125463
(182 digits)
Using MODMULN [mulredc:0, sqrredc:1]
Using B1=850000000, B2=15892628251516, polynomial Dickson(30), sigma=4292114945
dF=524288, k=5, d=5705700, d2=17, i0=132
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
15      47      168     661     2867    13623   69471   381778  2221086 1.4e+07
Step 1 took 5118300ms
Using 22 small primes for NTT
Estimated memory usage: 2153M
Initializing tables of differences for F took 6930ms
Computing roots of F took 87963ms
Building F from its roots took 42180ms
Computing 1/F took 15769ms
Initializing table of differences for G took 565ms
Computing roots of G took 68832ms
Building G from its roots took 43427ms
Computing roots of G took 68996ms
Building G from its roots took 44366ms
Computing G * H took 8462ms
Reducing  G * H mod F took 8758ms
Computing roots of G took 72232ms
Building G from its roots took 45154ms
Computing G * H took 8519ms
Reducing  G * H mod F took 8729ms
Computing roots of G took 72215ms
Building G from its roots took 46696ms
Computing G * H took 9187ms
Reducing  G * H mod F took 9270ms
Computing roots of G took 76390ms
Building G from its roots took 44273ms
Computing G * H took 8164ms
Reducing  G * H mod F took 8647ms
Computing polyeval(F,G) took 77520ms
Computing product of all F(g_i) took 330ms
Step 2 took 886306ms

[wombatman]

Carrying out 18k Stage 1 curves on C183_127_118.

[swellman]

Originally Posted by wombatman

I'll get started on the 18k curves needed at B1=110e6.

Factor found:

11:25:53 UTC Using B1=110000000-110000000, B2=776278396540, polynomial Dickson(30), sigma=3:2314982299
Sat 2016/07/30 11:25:53 UTC Step 1 took 0ms
Sat 2016/07/30 11:25:53 UTC Step 2 took 408988ms
Sat 2016/07/30 11:25:53 UTC ********** Factor found in step 2: 18741935848981531171662457713083107476972826270465020497
Sat 2016/07/30 11:25:53 UTC Found prime factor of 56 digits: 18741935848981531171662457713083107476972826270465020497
Sat 2016/07/30 11:25:53 UTC Prime cofactor ((97^128+128^97)/32654503840898125243550881908880946943525344683424710216799046973981058120030697)/18741935848981531171662457713083107476972826270465020497 has 120 digits

Thanks Ben! I'm starting Stage 2 on C179_136_87.

[swellman]

Originally Posted by wombatman

Preparing Stage 1 curves on C179_136_87 at B1=110e6.

Another hit!

Using B1=110000000-110000000, B2=776278396540, polynomial Dickson(30), sigma=3:4188775375
Tue 2016/08/02 07:49:28 UTC Step 1 took 0ms
Tue 2016/08/02 07:49:28 UTC Step 2 took 414089ms
Tue 2016/08/02 07:49:28 UTC ********** Factor found in step 2: 563655256452453087659366036867672874699806549622983
Tue 2016/08/02 07:49:28 UTC Found prime factor of 51 digits: 563655256452453087659366036867672874699806549622983
Tue 2016/08/02 07:49:28 UTC Prime cofactor ((87^136+136^87)/17772265331222093600920463956534435557326871232293582567248130632028948864220420721011)/563655256452453087659366036867672874699806549622983 has 128 digits

[wombatman]

Reserving C183_138_97 for Stage 1 work which will be sent on to Sean for Stage 2.

[swellman]

C179_148_87 is a composite cofactor (stub) from a yoyo ECM effort on C223_148_87. See http://factordb.com/index.php?id=1000000000044706172.

After ECM to t55, it should be a straightforward GNFS.

Ben - shall we enqueue this number for our Stage 1-2 process?

[wombatman]

Sure. I can throw that one on next. Does it just need the t55?