Factoring for a publication

[EdH]

Originally Posted by Max0526

EdH, please read through the explanation of what happens in CADO with SNFS polys:

Thanks! I wonder how I missed that - I read all the posts in these threads (I thought) thoroughly. . .

[Max0526]

Originally Posted by EdH

Thanks! I wonder how I missed that - I read all the posts in these threads (I thought) thoroughly. . .

Sorry, it was a PM.

[Max0526]

As I am planning to finalize the Magma script for the entire Figure 3 plot, I will need some minor help with all full stages 6-7 below the line 260 please.


EDIT: stage 4 and 5 are done in full already, so stages 6-7 please.

[Max0526]

The spin dropped it from SNFS 162 to SNFS 161.
The original best E was 3.25458293e-10 --> 4.19323182e-10 --> almost 29% up!
The E score is the same, order of coefficients and skew are different.
I am ECMing it at all different -work levels in yafu, it's past t30 so far.

Any takers?

Line 24 : {'c4': '8', 'c3': '4', 'c2': '1', 'c1': '2', 'c0': '2', 'y1': '5426010154312914975605874948738303532601', 'y0': '-8729190279481982682747568253526434138523'}
Skew, Murphy E:
0.94213 4.19323182e-10

Line 12 : {'c4': '2', 'c3': '2', 'c2': '1', 'c1': '4', 'c0': '8', 'y1': '8729190279481982682747568253526434138523', 'y0': '-5426010154312914975605874948738303532601'}
Skew, Murphy E:
1.06142 4.19323182e-10

[RichD]

Originally Posted by Max0526

The spin dropped it from SNFS 162 to SNFS 161.

Any takers?

I can take it. Please send the complete poly.

[Max0526]

Originally Posted by RichD

I can take it. Please send the complete poly.

(11, 0) c158 / snfs161
n: 84883988896532828882316343781284494349677398704711330183494412991591655822596309361577669265196914176022276366212763984772293268271740689501851622307308404009
Y0: -8729190279481982682747568253526434138523
Y1: 5426010154312914975605874948738303532601
c4: 8
c3: 4
c2: 1
c1: 2
c0: 2
skew: 0.94213
# E = 4.19323182e-10

Edit: RichD says on it.

[EdH]

Releasing line 161 (10, -8): All the small composites are factored, two of the three larger ones gave up small primes and the three larger composites that are left have all had t50 ECM done.

I'll move to line 162 and "play" there for a bit.

[bur]

Back to filtering again:

PID1369782 2021-06-14 15:38:46,206 Debug:Filtering - Duplicate Removal, removal pass: duplicates2.run(): Task state: {'stdiocount': 9, 'cputime_dup2': 396.8000000000029, 'realtime_dup2': 157.3701412677765, 'stats_dup2_time': '125.9', 'last_input_nrel': 23877961, 'last_output_nrel': 14106817, 'unique_ratio': 0.5768758743108373}
PID1369782 2021-06-14 15:38:57,308 Info:Filtering - Duplicate Removal, removal pass: 8112352 unique relations remain on slice 0
PID1369782 2021-06-14 15:39:08,691 Info:Filtering - Duplicate Removal, removal pass: 8114413 unique relations remain on slice 1
PID1369782 2021-06-14 15:39:08,691 Info:Filtering - Duplicate Removal, removal pass: Of 3570439 newly added relations 2119948 were unique (ratio 0.593750)
PID1369782 2021-06-14 15:39:08,692 Info:Filtering - Duplicate Removal, removal pass: 16226765 unique relations remain in total
PID1369782 2021-06-14 15:39:08,693 Info:Filtering - Singleton removal: Reading 16226765 unique and 1900439 free relations, total 18127204
PID1369782 2021-06-14 15:39:22,522 Info:Filtering - Singleton removal: After purge, 11467661 relations with 12650777 primes remain with excess -1183116
PID1369782 2021-06-14 15:39:22,522 Info:Filtering - Singleton removal: Not enough relations
PID1369782 2021-06-14 15:39:22,523 Info:Filtering - Singleton removal: Requesting 1183116 additional relations
PID1369782 2021-06-14 15:39:22,523 Info:Filtering - Duplicate Removal, removal pass: Got request for 17409881 (1183116 additional) output relations, estimate 29441016 (1992616 additional) needed in input

Does the last line mean, it needs 17409881 relations after singleton removal?

From 27000000 unfiltered relations came only 11467661 relations. If really 17.4e6 are required that would mean we're only 2/3 there - after 5 days. I guess at this point continuing is still better than restarting with different parameters though.

[charybdis]

Originally Posted by bur

Back to filtering again:
Does the last line mean, it needs 17409881 relations after singleton removal?

From 27000000 unfiltered relations came only 11467661 relations. If really 17.4e6 are required that would mean we're only 2/3 there - after 5 days. I guess at this point continuing is still better than restarting with different parameters though.

No, the figure of 17409881 just came from adding the deficit of 1183116 to the 16226765 existing unique relations. In other words, you need at least 1183116 more unique relations for filtering to succeed. In reality you will need more than that, because the number of primes in the relations will increase too. If required_excess is set to a positive value then you will need even more, but the resulting matrix will be smaller.

[bsquared]

Originally Posted by bur

Back to filtering again:
Does the last line mean, it needs 17409881 relations after singleton removal?

From 27000000 unfiltered relations came only 11467661 relations. If really 17.4e6 are required that would mean we're only 2/3 there - after 5 days. I guess at this point continuing is still better than restarting with different parameters though.

It looks like you have 16226765 unique (non-duplicate) relations. Out of 27M that's 40% duplication rate, which is high but not unprecedented. At this stage of the game, relations that survive singleton removal is not a linear function of input relations - relations will combine more and more readily as more are added. It looks like it is asking for 1992616 additional raw input relations, which is only 7% more than the 27M you've already gathered. I don't know if it accounts for the expected 40% duplication rate in that figure or not...

[bsquared]

Originally Posted by bsquared

I will work on line 128 (3,-9) next. At difficulty 221 and 222 both should be doable by SNFS if it comes to that.

I'm pushing toward a t60 on the c198 from line 128. Assuming the number survives that, I'll ask for a spun poly for the c198.