-   Aliquot Sequences (
-   -   Aliquot sequences that start on the integer powers n^i (

garambois 2023-02-25 06:52

Thank you both very much !
Me, I'm working on 2^577 and 2^583, as I'll still be able to run some CADO decompositions on the 128-threaded computer : my loan has been extended a bit.

VBCurtis 2023-02-25 16:14

Thanks, Ed! I'll decide this weekend whether to run it locally or feed it to NFS@home e-small queue. It'll be done quicker on e-small.

I'll run a little ECM on it, too, maybe 0.25 * t55

EdH 2023-02-25 18:13

[QUOTE=VBCurtis;625645]Thanks, Ed! I'll decide this weekend whether to run it locally or feed it to NFS@home e-small queue. It'll be done quicker on e-small.

I'll run a little ECM on it, too, maybe 0.25 * t55[/QUOTE]If you choose not to run it locally, would you want me to use adjusted params from the [URL=""]harvest thread[/URL] and run it? If I interpret the post correctly, the only changes from the c170 would be:[code]
tasks.lpb0 = 31
tasks.lpb1 = 31
tasks.sieve.mfb1 = 89[/code]

VBCurtis 2023-02-26 03:15

Sure, be my guest! You have more hardware than I, by about a factor of three. I'm presently entertained with this "get 13^i to 160 digits" mini-project, and C171 would be nearly 2 weeks for me.

I have a second params adjustment to suggest:
increase nq from 15625 to 78125, and decrease admax by a factor of 5. I'm finding better lognorm (stage 1) and poly scores at small ad-ranges, so searching more deeply down low seems profitable.

I'm still "in testing" with this adjustment, but for C100-115 it seems to produce polys 5-7% higher scoring, with a smaller advantage around C140 (only tested on 2 jobs so far at C14x).

EdH 2023-02-26 03:21

My cluster completed t55 without success. I will be interested in what size does show up. I'm not 100% confident in my cluster as this is the first successful completion at t55.

Additional info on how I got to t55:[code]
In addition to t20, t25, t30, t35, t40 runs,
2750 @ 11e6
4500 @ 43e6
4320 @ 26e7[/code]The last three calculated to 55.057 by

Sergiosi 2023-02-26 13:17

92^80 ends in a cycle

garambois 2023-02-26 15:23

And which cycle ?
The one of length 28 !
This had already happened only for 3^286, outside the bases themselves belonging to the C28 cycle !
This will be the only sequence ending with a cycle for base 92 as it was for base 3.
But for base 3, we were dealing with a sequence of opposite parity.

Did you Sergiosi terminate this sequence ?

Sergiosi 2023-02-26 20:59

Yes, I did the calculations for AS 92^80, starting at i4. The c144 there was the largest factorization needed for completing the sequence.

gd_barnes 2023-03-04 13:39

I have completed testing of new bases 106 & 108 all open sequences to >= 120 digits.

garambois 2023-03-04 15:36

Great job, Sergiosi and Gary.
Thank you very much to you.

gd_barnes 2023-03-06 00:32

48^5 merge info. has a typo. It shows as merged with sequence 15302802. It should be sequence 1530280.

48^5:i15 merges with 1530280:i2

A point of interest: 15302802 does merge with the main project:
15302802:i189 merges with 27860:i19.

I also found it strange that 31^36 shows as merged with sequence 3762570. I verified that the merge info. is correct but I didn't know you were showing merges with sequences outside of the main project. I confirmed that neither 31^36 nor 3762570 merge with any sequences < 3M.

I only caught these due to the new modifications to the page to include new data files by Karsten.

All times are UTC. The time now is 10:11.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, Jelsoft Enterprises Ltd.