Reserved for MF - Sequence 4788 - Page 245

Raman · 2016-09-25, 09:38

Quote:

Originally Posted by science_man_88

now has 2^10 on the last iteration shown.

Not only that out.

Did any one else certainly notice away with in this unusual similarity only up?

Aliquot Sequence 660 Iteration Number 971 = 2⁵ × 2069 × c193.

Aliquot Sequence 4788 Iteration Number 10616 = 2¹⁰ × 2069 × c175.

GCD(Aliquot Sequence 660 Iteration Number 971, Aliquot Sequence 4788 Iteration Number 10616) = What?

Quote:

Originally Posted by Raman

@ Ryan Propper: Have you automated submissions to FactorDB even when you are asleep or outside? Great job!
Factoring a c120 in 2006 would take upto 5 days. In 2016, it is only taking 5 minutes. Or 10 minutes may be? Distributed computing is being going on automatically?
Perhaps you could help out with aliquot sequences of 552, 564, 660, 1512, 1992, 5250, 9120, 11040 besides of 4788 or of 314718.

<snip>

away out off up down my own - that ever which ever a way a way ever.
away out off up down my own - that ever which ever a way a way ever.

Aliquot Sequence 4788 Iteration Number 10615: 2⁶ × 113 - hopefully - not - 2⁶ × 127!

Consecutive Prime Numbers 113 And 127!

science_man_88 · 2016-09-25, 13:19

Quote:

Originally Posted by Raman

Not only that out.

Did any one else certainly notice away with in this unusual similarity only up?

Aliquot Sequence 660 Iteration Number 971 = 2⁵ × 2069 × c193.

Aliquot Sequence 4788 Iteration Number 10616 = 2¹⁰ × 2069 × c175.

GCD(Aliquot Sequence 660 Iteration Number 971, Aliquot Sequence 4788 Iteration Number 10616) = What?

Aliquot Sequence 4788 Iteration Number 10615: 2⁶ × 113 - hopefully - not - 2⁶ × 127!

Consecutive Prime Numbers 113 And 127!

using PARI/GP to try to find the gcd of the composites that haven't been factored
( after reducing it enough I tried gcd of the two last results and got 1. so $2^5\cdot2069$ edit: okay I used iteration 971 for 660 but the point is the same.

ryanp · 2016-09-27, 17:10

Plodding along slowly. That 2^2 (and now 2^6) is hard to shake...

schickel · 2016-09-27, 23:49

Amazing, I was just re-reading this thread that showed up in the "Similar threads" section, and 5 years ago we took ~2 weeks getting ready for a c172. Now we're having this happen:

Quote:

Originally Posted by ryanp;432854 (on 4/30)

I'll handle the next C180 as well. :)

Quote:

Originally Posted by ryanp;432959 (on 5/2)

OK. Next C180 is done:

Code:

Mon May  2 13:11:50 2016  prp63 factor: 852828185224582024294795864244283338007551456801212478326125773
Mon May  2 13:11:50 2016  prp117 factor: 704131409692002901688336495667355706435373199424596628222345220771103914385162349468158740451400667823527752200435477

And

Quote:

Originally Posted by ryanp;432991 (again, on 5/2)

Working on the C170 now.

Code:

linear algebra completed 733239 of 5927756 dimensions (12.4%, ETA 10h 9m)

Quote:

Originally Posted by ryanp;433019 (on the next day, 5/3)

Awww. Could've had this one by ECM if I had tried harder...

Code:

Tue May  3 08:55:24 2016  prp50 factor: 10343669734301041373937055383512171230895724608491
Tue May  3 08:55:24 2016  prp120 factor: 993739759889626805545135903141953923233113832717632966003694092677384994673520180146397186251215841015082033396156750101

We just need ryanp to get mad at all the sequences!

Thanks for the yuuge assist, Ryan!!

LaurV · 2016-09-28, 01:58

Quote:

Originally Posted by schickel

We just need ryanp to get mad at all the sequences!

Well... I already made a diabolic plan: I am waiting for this sequence to get to 202 digits (and hopefully with 2^3*3*5 driver

), and then I will quote Ryan's post where he said that "he is pissed off" and "he will kill it", and I will ask "did you say something?"

(well, you all jinxed it that it will terminate, and it didn't, so I try the other way around... hehe)

ryanp · 2016-09-28, 02:00

Quote:

Originally Posted by LaurV

Well... I already made a diabolic plan: I am waiting for this sequence to get to 202 digits (and hopefully with 2^3*3*5 driver

), and then I will quote Ryan's post where he said that "he is pissed off" and "he will kill it", and I will ask "did you say something?"

I'll find a way...

flagrantflowers · 2016-09-28, 03:44

Quote:

Originally Posted by ryanp

I'll find a way...

Just so that things are jinxed in a clearly defined way; what is your limit, in terms of factor size? C200? C210?

legendarymudkip · 2016-09-28, 15:01

Quote:

Originally Posted by flagrantflowers

Just so that things are jinxed in a clearly defined way; what is your limit, in terms of factor size? C200? C210?

In September 2013, he factored RSA210. Given that was (almost exactly) 3 years ago, assuming a rate of Moore's Law at 18 months, he should be able to factor a C220 by now. From this it should be at least C220!.

VBCurtis · 2016-09-28, 16:45

Such extrapolation works fine for the sieving step, but the hurdle to really big jobs is handling the matrix. Clusters may have gotten faster in 3 years, but cluster time likely hasn't gotten easier to find. RSA numbers are interesting enough to make a case for time on a nice cluster, but "I want to kill this %$^&%ing Aliqueit sequence" might not gain similar access.

On the other hand, 64GB memory desktops are now available, and a year on one of those might solve a GNFS-215 matrix without a cluster. So, perhaps you're right about 215+, given enough patience.

RichD · 2016-09-28, 20:18

So is it safe to say, "You can't over-sieve your way to less memory requirements."

GP2 · 2016-09-28, 21:28

Quote:

Originally Posted by VBCurtis

On the other hand, 64GB memory desktops are now available, and a year on one of those might solve a GNFS-215 matrix without a cluster. So, perhaps you're right about 215+, given enough patience.

In the cloud you can get 61 GiB of memory on a 4-core Xeon @ 2.5 GHz at about $1000 for a year's worth of computing time at the cheapest current spot prices. You can get 244 GiB of memory on a 16-core for maybe three times that amount.

The drawback is, spot prices fluctuate, so much of the time the spot price would exceed what you'd want to pay. So a year's worth of computing time might be way longer in wall-clock time...

Edit: Google cloud prices are similar, 4-core 52 GB preemptible @ $73/month or 16-core 208 GB preemptible @ $292/month, nonfluctuating prices and maybe a better chance of getting interrupted less often.

Are there any interesting problems that need humungous memory but considerably less than a year to solve?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Reserved for MF - Sequence 3366	RichD	Aliquot Sequences	470	2021-04-22 02:17
Reserved for MF - Sequence 3408	RichD	Aliquot Sequences	474	2021-03-07 20:28
Reserved for MF - Sequence 276	kar_bon	Aliquot Sequences	127	2020-12-17 10:05
Assignments are reserved but not showing up	prism019	GPU to 72	6	2020-09-21 22:11
80M to 64 bits ... but not really reserved	petrw1	Lone Mersenne Hunters	82	2010-01-11 01:57

2016-09-27, 17:10	#2687
ryanp Jun 2012 Boulder, CO 17² Posts	Plodding along slowly. That 2^2 (and now 2^6) is hard to shake...

2016-09-28, 16:45	#2693
VBCurtis "Curtis" Feb 2005 Riverside, CA 11375₈ Posts	Such extrapolation works fine for the sieving step, but the hurdle to really big jobs is handling the matrix. Clusters may have gotten faster in 3 years, but cluster time likely hasn't gotten easier to find. RSA numbers are interesting enough to make a case for time on a nice cluster, but "I want to kill this %$^&%ing Aliqueit sequence" might not gain similar access. On the other hand, 64GB memory desktops are now available, and a year on one of those might solve a GNFS-215 matrix without a cluster. So, perhaps you're right about 215+, given enough patience.

2016-09-28, 20:18	#2694
RichD Sep 2008 Kansas 6460₈ Posts	So is it safe to say, "You can't over-sieve your way to less memory requirements."