GNFS targets which need more ECM

[Max0526]

@fivemack

Don't worry about the polynomial, it was just a baseline. Don't forget to update http://factordb.com/index.php?id=1100000000408940052.

[swellman]

Originally Posted by fivemack

I will now join swellman with an initial 12800 curves at 3e8 on C197_141_86.

I'm closing in on 10,000 curves total @B1=3e8. Should reach the 10k mark in approximately a week. My CPUs are getting a workout!

Fivemack - do you plan on running a second batch of 12800 curves (if necessary) to complete ECM? Or should I plan on doing a few more curves? Will help however I can but my CPU throughout is a fraction of your GPUs.

[Max0526]

If it comes to GNFS:

Msieve v. 1.52 (SVN 958)
R0: -130146348118863904411208919949896430
R1: 2012808155599133609497
A0: 864426134371135342504339042798381433197179
A1: 1016050927402686276428373958463009017
A2: 693072552180938852726455110
A3: -7434828005701894904041
A4: 19061856712467
A5: 2859300
skew 23156229.62, size 8.173e-018, alpha -6.331, combined = 6.371e-014

Found by CADO-NFS.

[Max0526]

Found by CADO-NFS, polished by Msieve:

Msieve v. 1.52 (SVN 958)
R0: -130438779319684744675863928393114965
R1: 102279529274362597738967
A0: -84316679322238097560554777104612811139890544
A1: 42403409232487736051929317973702780684
A2: 320298423355775483396313447392
A3: -26568518681697447959709
A4: -74324527387088
A5: 3455760
skew 57479716.34, size 8.271e-018, alpha -7.851, combined = 6.316e-014

[fivemack]

Originally Posted by swellman

I'm closing in on 10,000 curves total @B1=3e8. Should reach the 10k mark in approximately a week. My CPUs are getting a workout!

Fivemack - do you plan on running a second batch of 12800 curves (if necessary) to complete ECM? Or should I plan on doing a few more curves? Will help however I can but my CPU throughout is a fraction of your GPUs.

It looks as if you're running about 10,000 curves a month; so if you plan to do another 5,000, and I carry on up to 20480 total on the GPU, we get to 35K around the start of November. I'm happy with that as a strategy unless you have something more interesting to do with your CPUs.

[swellman]

Originally Posted by fivemack

It looks as if you're running about 10,000 curves a month; so if you plan to do another 5,000, and I carry on up to 20480 total on the GPU, we get to 35K around the start of November. I'm happy with that as a strategy unless you have something more interesting to do with your CPUs.

Sure, I'll plan on running another 5,000 curves. Should be done by late Oct.

[amphoria]

Originally Posted by amphoria

I also ran 2.5k curves at B1=300M but do not intend to run anymore at this time.

I am running some more curves at B1=300M.

[amphoria]

Originally Posted by amphoria

I am running some more curves at B1=300M.

I have now run 12k curves at B1=300M (including the original 2.5k) with no factors found.

Whilst the CPU was running stage 2 on the final batch, I ran msieve poly select on the GPU. The best poly found (vs an msieve target of 3.66e-014 to 4.21e-014+) was

# norm 4.544455e-018 alpha -7.010993 e 3.925e-014 rroots 5
skew: 42379359.70
c0: -124664909780140586713358287271932757503367520
c1: 19857214532726171166338731698655435478
c2: -80374314450663677283618496297
c3: -37388316203874674366923
c4: -86114669846817
c5: 4697244
Y0: -528584393658161950693242720222768713
Y1: 17611375245262387

[swellman]

Originally Posted by amphoria

I have now run 12k curves at B1=300M (including the original 2.5k) with no factors found.

Whilst the CPU was running stage 2 on the final batch, I ran msieve poly select on the GPU. The best poly found (vs an msieve target of 3.66e-014 to 4.21e-014+) was

# norm 4.544455e-018 alpha -7.010993 e 3.925e-014 rroots 5
skew: 42379359.70
c0: -124664909780140586713358287271932757503367520
c1: 19857214532726171166338731698655435478
c2: -80374314450663677283618496297
c3: -37388316203874674366923
c4: -86114669846817
c5: 4697244
Y0: -528584393658161950693242720222768713
Y1: 17611375245262387

The best SNFS poly I could find for this composite (on 15e siever):

n: 193827236134443466098224493848711326783427017152262904952196883820082655569819229847578714306889013642271468466999382170099561997247424941621375124328481656372181577090568838136723217969
# 145^72+72^145, difficulty: 271.72, anorm: 1.70e+037, rnorm: -7.61e+050
# scaled difficulty: 274.00, suggest sieving rational side
# size = 2.209e-013, alpha = 0.000, combined = 3.756e-014, rroots = 0
type: snfs
size: 271
skew: 2.040
c6: 1
c0: 72
Y1: 376686377243952195401306999861472444605792256
Y0: -86380562306022715087890625
rlim: 268000000
alim: 268000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Test sieving indicates the SNFS is a bit faster with a slightly higher yield. Continuing to search for a better GNFS poly could change this result but it would have to be a big leap in quality.

[swellman]

Originally Posted by fivemack

It looks as if you're running about 10,000 curves a month; so if you plan to do another 5,000, and I carry on up to 20480 total on the GPU, we get to 35K around the start of November. I'm happy with that as a strategy unless you have something more interesting to do with your CPUs.

Completed 15000 curves @B1=3e8 with no factors found.

[fivemack]

C197_141_86: 20480 curves at B1=3e8 completed with no factor found; so that's the 35K done.

Shall I set the GTX970 to run polynomial search for two weeks?