Treading LL P-1 for 100M TF - Page 25

Aramis Wyler · 2013-10-08, 13:09

The last post is here, with 3369.6 GHD/10. There's been 5 factors so far, including todays set. GHD/10 so far is now up to 4348.45.

Code:

Number     |  Work   |  Range  |  Date Completed    |   GHzD   |  Status
332244137  |  LL TF  |  75-79  |  2013-10-08 06:44  |  345.47  |  No Factor
332244091  |  LL TF  |  75-79  |  2013-10-08 03:43  |  345.47  |  No Factor
332243309  |  LL TF  |  76-79  |  2013-10-07 06:44  |  322.44  |  No Factor
332243287  |  LL TF  |  76-79  |  2013-10-07 01:43  |  322.44  |  No Factor
332243279  |  LL TF  |  75-79  |  2013-10-06 08:45  |  345.47  |  No Factor
332231213  |  LL TF  |  77-79  |  2013-10-06 01:44  |  276.39  |  No Factor
332243209  |  LL TF  |  75-79  |  2013-10-05 08:43  |  345.47  |  No Factor
332243203  |  LL TF  |  75-79  |  2013-10-05 04:43  |  345.47  |  No Factor
332243159  |  LL TF  |  75-79  |  2013-10-04 09:45  |  345.47  |  No Factor
332243099  |  LL TF  |  75-79  |  2013-10-04 02:44  |  345.47  |  No Factor
332243029  |  LL TF  |  75-79  |  2013-10-03 09:43  |  345.47  |  No Factor
332243003  |  LL TF  |  75-79  |  2013-10-03 01:43  |  345.47  |  No Factor
332242969  |  LL TF  |  75-79  |  2013-10-02 11:13  |  345.47  |  No Factor
332240159  |  LL TF  |  74-79  |  2013-10-01 23:43  |  356.99  |  No Factor
332240089  |  LL TF  |  74-79  |  2013-10-01 11:43  |  356.99  |  No Factor
332239253  |  LL TF  |  74-79  |  2013-09-30 21:44  |  356.99  |  No Factor
332238437  |  LL TF  |  74-79  |  2013-09-30 11:43  |  356.99  |  No Factor
332238407  |  LL TF  |  74-79  |  2013-09-29 19:43  |  356.99  |  No Factor
332238073  |  LL TF  |  74-79  |  2013-09-29 11:43  |  356.99  |  No Factor
332237981  |  LL TF  |  74-79  |  2013-09-28 16:44  |  357.00  |  No Factor
332237797  |  LL TF  |  74-79  |  2013-09-28 10:43  |  357.00  |  No Factor
332237483  |  LL TF  |  75-79  |  2013-09-27 14:43  |  345.48  |  No Factor
332237617  |  LL TF  |  74-79  |  2013-09-27 10:43  |  357.00  |  No Factor
332234827  |  LL TF  |  74-79  |  2013-09-26 12:43  |  357.00  |  No Factor
332237453  |  LL TF  |  74-79  |  2013-09-26 10:43  |  357.00  |  No Factor
332232931  |  LL TF  |  74-79  |  2013-09-25 10:43  |  357.00  |  No Factor
332232977  |  LL TF  |  74-79  |  2013-09-25 10:43  |   80.61  |  118787691707013852544999
332232949  |  LL TF  |  74-79  |  2013-09-25 07:43  |  357.00  |  No Factor
332242957  |  LL TF  |  75-79  |  2013-09-24 09:43  |  345.47  |  No Factor

kladner · 2013-10-09, 03:00

[QUOTE=kladner;355567]I have

Code:

Factor=N/A,332244xxx,75,79 
Factor=N/A,332244xxx,75,79

in queue.

They will be finished about 0700 CDT.

kladner · 2013-10-13, 06:59

A few more-

Code:

332244553    LL TF    75    79    2013-10-08 01:25:32    2013-10-09 16:44:42    345.472
332244569    LL TF    75    79    2013-10-08 01:25:32    2013-10-09 16:44:42    345.472
332284207    LL TF    74    79    2013-10-08 01:41:49    2013-10-13 06:45:03    356.945

kladner · 2013-10-14, 22:29

One more-

Code:

332284207    LL TF    74    79     2013-10-13 356.945 GHz-days

kladner · 2013-10-15, 20:53

And another:

Quote:

no factor for M332246729 from 2^75 to 2^76 [mfaktc 0.20 barrett76_mul32_gs] CPU credit is 23.0313 GHz-days.
no factor for M332246729 from 2^76 to 2^77 [mfaktc 0.20 barrett77_mul32_gs] CPU credit is 46.0627 GHz-days.
no factor for M332246729 from 2^77 to 2^78 [mfaktc 0.20 barrett87_mul32_gs] CPU credit is 92.1253 GHz-days.
no factor for M332246729 from 2^78 to 2^79 [mfaktc 0.20 barrett87_mul32_gs] CPU credit is 184.2507 GHz-days

kladner · 2013-10-22, 04:58

Completed an hour ago-

Quote:

332247761 - LLTF 75-79 2013-10-22 - 345.468

kladner · 2013-10-25, 04:55

332265467 _ LL TF_ 75-79 2013-10-24 00:35:21 2013-10-25 02:43:41 345.450

I am curious why exponents factored to 79 do not then appear in that column in the GPU72 Available TF table. These seem to be getting released to LL workers. Is there no provision for P-1 work**, or is this an inefficient use of processing power in this range?

**EDIT: .....or further TF work....

Aramis Wyler · 2013-10-26, 16:19

Any numbers TF'd to 79 or higher are released for work. It used to be lower, but now 79 is the number. You can factore numbers tohigher than 79 through gpu72, but you can't start at 79. It is possible to start at 78 if you use the 'highest factored' option.

Regarding the efficiency of factoring higher than 79, I think that is the case on only a couple of cards. Personally I don't like factoring higher than 79 because it messes with my numbers - I can factor from 64 to 79 in 24 hours. But going from 79 to 80 takes 25 all by itself, due to the mfactc algorithms getting less efficient past 76.

kladner · 2013-10-26, 17:01

Quote:

Originally Posted by Aramis Wyler

[snip]
Regarding the efficiency of factoring higher than 79, I think that is the case on only a couple of cards. Personally I don't like factoring higher than 79 because it messes with my numbers - I can factor from 64 to 79 in 24 hours. But going from 79 to 80 takes 25 all by itself, due to the mfactc algorithms getting less efficient past 76.

Exactly. Each step higher seems to take a bit longer than all that have gone before combined. I don't mean this as any exact representation, but I've seen a similar proportion.

Thanks for the answer. I was just wondering.

Am I presenting results here in useful format? I'm not really keeping track, but I know you were maintaining a spreadsheet balancing this work with Uncwilly's P-1 output. If a particular set of numbers would be helpful, i.e. the Manual Results output page, I'll make more of a point of doing it that way.

Aramis Wyler · 2013-10-27, 18:21

I tend to post mine in batches of thousands of hours, to make tracking them easy. I don't think the format much matters, I only use the format I do because it's easy to modify gpu72's output to display it. Since only 1/10th the GHDs we put out are to match Uncwilly's GHD, putting a given number out here from 74-79 is what, 341 GHDs or so? So it matches to 34 of Unc's. Since he's got his minions running p-1 at record setting rates, we need lots of those. So yeah, my wrapups are just so I don't have to post them 4 times a day. :-D

I haven't done yours or Kracker's since... I don't even remember - August? Because they've been coming out in small pieces. Maybe I'll parse through the forums for the past couple months and roll them up. They are certainly appreciated.

And certainly, if you or anyone can come up with a better way to track the numbers I'm all ears. Unc's numbers are easy to track because he uses gpu72 exclusively for this - I can just glance at his red line to see where he is. There's no easy way for me to separate my LMH tf from my regular tf on the public graph.

chalsall · 2013-10-27, 18:24

Quote:

Originally Posted by Aramis Wyler

And certainly, if you or anyone can come up with a better way to track the numbers I'm all ears. Unc's numbers are easy to track because he uses gpu72 exclusively for this - I can just glance at his red line to see where he is. There's no easy way for me to separate my LMH tf from my regular tf on the public graph.

I'm still trying to keep my head above water with regards to "real-work", but when I have some cycles I'll add the ability to select one's "Completed" report results based on range and/or dates.

2013-10-09, 03:00	#266
kladner "Kieren" Jul 2011 In My Own Galaxy! 2×3×1,693 Posts	[QUOTE=kladner;355567]I have Code: Factor=N/A,332244xxx,75,79 Factor=N/A,332244xxx,75,79 in queue. They will be finished about 0700 CDT.

2013-10-14, 22:29	#268
kladner "Kieren" Jul 2011 In My Own Galaxy! 2·3·1,693 Posts	One more- Code: 332284207 LL TF 74 79 2013-10-13 356.945 GHz-days

2013-10-25, 04:55	#271
kladner "Kieren" Jul 2011 In My Own Galaxy! 2×3×1,693 Posts	332265467 _ LL TF_ 75-79 2013-10-24 00:35:21 2013-10-25 02:43:41 345.450 I am curious why exponents factored to 79 do not then appear in that column in the GPU72 Available TF table. These seem to be getting released to LL workers. Is there no provision for P-1 work, or is this an inefficient use of processing power in this range? EDIT: .....or further TF work.... Last fiddled with by kladner on 2013-10-25 at 04:56

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
To 100M to 75 Bits	petrw1	GPU to 72	6	2017-04-17 04:03
Is LMH > 100M still alive?	TheMawn	LMH > 100M	25	2016-04-18 16:11
100m p-1 and tf	aurashift	Software	18	2016-04-14 13:48
Anyone working in 79.3M - 100M ??	markr	Lone Mersenne Hunters	21	2008-12-21 16:02
100M madness	stars10250	Hardware	8	2008-10-02 15:21

2013-10-26, 16:19	#272
Aramis Wyler "Bill Staffen" Jan 2013 Pittsburgh, PA, USA 1A8₁₆ Posts	Any numbers TF'd to 79 or higher are released for work. It used to be lower, but now 79 is the number. You can factore numbers tohigher than 79 through gpu72, but you can't start at 79. It is possible to start at 78 if you use the 'highest factored' option. Regarding the efficiency of factoring higher than 79, I think that is the case on only a couple of cards. Personally I don't like factoring higher than 79 because it messes with my numbers - I can factor from 64 to 79 in 24 hours. But going from 79 to 80 takes 25 all by itself, due to the mfactc algorithms getting less efficient past 76.

2013-10-27, 18:21	#274
Aramis Wyler "Bill Staffen" Jan 2013 Pittsburgh, PA, USA 2³·53 Posts	I tend to post mine in batches of thousands of hours, to make tracking them easy. I don't think the format much matters, I only use the format I do because it's easy to modify gpu72's output to display it. Since only 1/10th the GHDs we put out are to match Uncwilly's GHD, putting a given number out here from 74-79 is what, 341 GHDs or so? So it matches to 34 of Unc's. Since he's got his minions running p-1 at record setting rates, we need lots of those. So yeah, my wrapups are just so I don't have to post them 4 times a day. :-D I haven't done yours or Kracker's since... I don't even remember - August? Because they've been coming out in small pieces. Maybe I'll parse through the forums for the past couple months and roll them up. They are certainly appreciated. And certainly, if you or anyone can come up with a better way to track the numbers I'm all ears. Unc's numbers are easy to track because he uses gpu72 exclusively for this - I can just glance at his red line to see where he is. There's no easy way for me to separate my LMH tf from my regular tf on the public graph.