Some Somewhat Easier n^i Sequences Available for Termination - Page 21

gd_barnes · 2022-05-25, 20:10

Quote:

Originally Posted by EdH

I ran the c159 (about 26 hours) <snip>

26 hours for a C159. Wow! How many cores on that machine? Either I'm doing something wrong or that is a big machine! :-)

VBCurtis · 2022-05-25, 20:31

Ed's farm is something like 100 cores spread over a ton of rather old hardware, like HP Z600 age or even some Core2.
His magic is in the scripting to get them all to work together. :)

My notes only have one C159 job. It took 3.2 days as the only thing running on a Z620 dual-10-core Ivy Bridge Xeon.
I just got a 5950X; haven't started CADO timing runs yet, too giddy with LLR timings still.

EdH · 2022-05-26, 03:28

Quote:

Originally Posted by gd_barnes

26 hours for a C159. Wow! How many cores on that machine? Either I'm doing something wrong or that is a big machine! :-)

I've actually got about 250 threads (across a lot of machines - about a third are i7s) running the CADO-NFS polyselect and sieving (a few less during polyselect), and then Msieve is allowed 40 threads for LA and 8 for SR. LA takes quite a bit longer than sieving, but the clients are freed to move to another task while LA and SR run.

gd_barnes · 2022-05-26, 07:14

Quote:

Originally Posted by EdH

I've actually got about 250 threads (across a lot of machines - about a third are i7s) running the CADO-NFS polyselect and sieving (a few less during polyselect), and then Msieve is allowed 40 threads for LA and 8 for SR. LA takes quite a bit longer than sieving, but the clients are freed to move to another task while LA and SR run.

Very impressive! I feel much better now.

I've been running everything for factoring on an 8-core/8-thread I7-9700 and I'm pretty happy to be able to ECM/sieve a C100 in 20 mins or a C120 in ~2.5 hours. The biggest I have is a Ryzen 3950X...16 cores/32 threads but I like to have it for sieving and prime searching and occasional small factoring jobs.

The magic really is in that scripting! That's a lot of threads to bring together.

gd_barnes · 2022-05-26, 09:52

38^100 terminates

VBCurtis · 2022-05-26, 16:10

Quote:

Originally Posted by gd_barnes

I've been running everything for factoring on an 8-core/8-thread I7-9700 and I'm pretty happy to be able to ECM/sieve a C100 in 20 mins or a C120 in ~2.5 hours.

Is this CADO or GGNFS/yafu?

That same Z620 has 8min 20 sec for a C103 and almost exactly 60 min for C120, running CADO with the params files posted to the CADO forum. If you're in Linux, I strongly suggest getting CADO set up with the updated params files. I think the 9700 is about twice as fast per-core as my Z620 on CADO, so your 8 cores should achieve 75-80% of the speed of my Z620.

Or maybe you have other tasks running and my speed comparisons are folly.

gd_barnes · 2022-05-26, 21:03

Quote:

Originally Posted by VBCurtis

Is this CADO or GGNFS/yafu?

That same Z620 has 8min 20 sec for a C103 and almost exactly 60 min for C120, running CADO with the params files posted to the CADO forum. If you're in Linux, I strongly suggest getting CADO set up with the updated params files. I think the 9700 is about twice as fast per-core as my Z620 on CADO, so your 8 cores should achieve 75-80% of the speed of my Z620.

Or maybe you have other tasks running and my speed comparisons are folly.

No other tasks running.

I'm in Windows using GGNFS/yafu. All of my modern machines run Windows. Only an old server machine is running Linux.

I'm not familiar with CADO. Can it be used in Windows?

gd_barnes · 2022-05-26, 23:11

15472^36 terminates

gd_barnes · 2022-05-27, 03:35

30030^34 terminates

gd_barnes · 2022-05-27, 04:55

103^77 and 119^75 terminate

gd_barnes · 2022-05-27, 06:54

59^87 and 193^59 terminate

2022-05-27, 04:55	#230
gd_barnes May 2007 Kansas; USA 11115₁₀ Posts	103^77 and 119^75 terminate Last fiddled with by gd_barnes on 2022-05-27 at 05:35

2022-05-27, 06:54	#231
gd_barnes May 2007 Kansas; USA 10101101101011₂ Posts	59^87 and 193^59 terminate Last fiddled with by gd_barnes on 2022-05-27 at 07:51

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Unexpected termination of PM-1	Miszka	Software	22	2021-11-19 21:36
Easier pi(x) approximation	mathPuzzles	Math	8	2017-05-04 10:58
Would finding a definate Pi value easier if...	xtreme2k	Math	34	2013-09-09 23:54
Aliquot Termination Question - Largest Prime?	EdH	Aliquot Sequences	6	2010-04-06 00:12
A new termination below 100k	10metreh	Aliquot Sequences	0	2010-03-11 18:24

2022-05-25, 20:31	#222
VBCurtis "Curtis" Feb 2005 Riverside, CA 5378₁₀ Posts	Ed's farm is something like 100 cores spread over a ton of rather old hardware, like HP Z600 age or even some Core2. His magic is in the scripting to get them all to work together. :) My notes only have one C159 job. It took 3.2 days as the only thing running on a Z620 dual-10-core Ivy Bridge Xeon. I just got a 5950X; haven't started CADO timing runs yet, too giddy with LLR timings still.

2022-05-26, 09:52	#225
gd_barnes May 2007 Kansas; USA 3²×5×13×19 Posts	38^100 terminates

2022-05-26, 23:11	#228
gd_barnes May 2007 Kansas; USA 25553₈ Posts	15472^36 terminates

2022-05-27, 03:35	#229
gd_barnes May 2007 Kansas; USA 3²×5×13×19 Posts	30030^34 terminates