Full report for 2006 year

Yxine · 2007-01-04, 08:30

I generated the usual year report of done work:
http://zip.larin.name/MasterPrime2006-Stats.zip (1.183.606 bytes)
Also, here is a full statuses reports (113 reports for 11 years):
http://zip.larin.name/MasterPrime2006-Statuses.zip (581.367 bytes)

jinydu · 2007-01-04, 09:57

Here is a question I am curious about: What is the total amount of work done by GIMPS in its entire lifetime so far? The estimate given at http://mersenne.org/ips/stats.html is 11 months out of date.

Yxine · 2007-01-04, 15:08

I can't count this... sorry!

S485122 · 2007-01-04, 15:28

On the Prime Net Top Producer page, one can get the full report Top Producers Awards, Full Length Report

As of 2007-01-04 14:00 UTC the totals where

Code:

LL P90         Exponents  Fact.P90    Exponents  P90 CPU hrs/day
CPU yrs        LL Tested  CPU yrs     w/ Factor  
1 958 357,026  888 875    60 071,380  103 786    15 202 356,250

!!! I use the ISO character for the decimal separator : the comma.

(Lazily computed by totalling in Excel after breaking the report over two sheets.)

jinydu · 2007-01-05, 03:13

Thanks for the figures S485122. So GIMPS has indeed passed the 2 million CPU years mark. It would interesting to see how that figure has changed over the course of months and years.

By the way, if all you did was add all the entries together, the P90 CPU hrs/day figure is incorrect. For instance, I have 3094.60 CPU hrs/day. This is my average speed since joining on Nov. 18, 2003; not my average speed since the start of GIMPS, which is obviously several times lower. So the figure of 15202356.250 CPU hrs/day is definitely an overestimate.

Also, sorry for being nitpicky, but even the figures for LL P90 CPU years and Factoring P90 CPU years are not exactly correct because some GIMPS participants do not use Primenet. Still, I would think that it is a good approximation.

S485122 · 2007-01-05, 06:39

Quote:

Originally Posted by jinydu

Thanks for the figures S485122. So GIMPS has indeed passed the 2 million CPU years mark. It would interesting to see how that figure has changed over the course of months and years.

Unfortunately the relevant file is not archived periodically.

Quote:

Originally Posted by jinydu

By the way, if all you did was add all the entries together, the P90 CPU hrs/day figure is incorrect. For instance, I have 3094.60 CPU hrs/day. This is my average speed since joining on Nov. 18, 2003; not my average speed since the start of GIMPS, which is obviously several times lower. So the figure of 15202356.250 CPU hrs/day is definitely an overestimate.

Indeed, I should have left the CPU hours a day out.

Quote:

Originally Posted by jinydu

Also, sorry for being nitpicky, but even the figures for LL P90 CPU years and Factoring P90 CPU years are not exactly correct because some GIMPS participants do not use Primenet. Still, I would think that it is a good approximation.

This is why I cited my "sources" :-) The problem with the GIMPS statistics is that one has to scan 6 files and that factoring is not tallied. This is why I choose to be lazy and use the PrimeNet page. This afternoon I will do the same with the GIMPS statistics. It would allow us to compare both statistics.

Jacob

davieddy · 2007-01-05, 12:39

Quote:

Originally Posted by Yxine

I generated the usual year report of done work:
http://zip.larin.name/MasterPrime2006-Stats.zip (1.183.606 bytes)
Also, here is a full statuses reports (113 reports for 11 years):
http://zip.larin.name/MasterPrime2006-Statuses.zip (581.367 bytes)

Thankyou very much. Entertaining and informative.

I especially liked the graphs as an overall picture.
Excel has done some smoothing I guess.
For example the once-LLtested graph should jump
vertically at exponent 33,219,281 reflecting the
effort to find a 10,000,000 digit prime.

David

S485122 · 2007-01-05, 12:47

As promised the total from the GIMPS pages.
Since the files on ftp://mersenne.org/gimps/ are dated from 2007-01-01 16:10 to 16:54 that must be the date of the GIMPS Top Producers statistics as well.

Code:

Who		      P-90 CPU years  Exponents tested
Total      1-   100	  649 331,69	  332 470
Total    101-10 000	1 224 169,41	  688 741
Total 10 001-20 000	  123 557,42	  106 847
Total 20 001-30 000	   42 975,92	   51 234
Total 30 001-40 000	   15 174,87	   51 813
Total 40 001-50 565	    3 475,56	   36 656
Grand Total 2007-01-01	2 058 684,87	1 267 761

The difference with PrimeNet is big indeed.

Code:

Grand Total             1 958 357,026	  888 875

jinydu · 2007-01-05, 13:11

How did you determine the number of CPU years completed? If I am not mistaken, only the number of CPU years remaining is shown.

Also, what about the factoring CPU years completed.

Good effort though, in any case.

S485122 · 2007-01-05, 13:43

Quote:

Originally Posted by jinydu

How did you determine the number of CPU years completed? If I am not mistaken, only the number of CPU years remaining is shown.

I just went to the Top Producers page, copied all the data drom the different groups and added them up. That page provides the credit for LL work completed, not for the work to do.

Quote:

Originally Posted by jinydu

Also, what about the factoring CPU years completed.

The page I got my results from says :

Quote:

This page lists the top producers of Lucas-Lehmer test results sorted by Pentium-90 CPU-years. First, some caveats. I do not track the amount of time spent in factoring Mersenne numbers. The number of CPU years is based on version 19 of the program. This understates the actual time because older program versions were less efficient. This also understates the efforts of UNIX and Mac users as the programs available on those platforms are not as efficient as the Pentium version. Finally, if a factor is later found for a Mersenne number or the Lucas-Lehmer result is found to be incorrect, then you will "lose credit" for the time spent running the test.

So I could not compute factoring times because I do not have the data.

Quote:

Originally Posted by jinydu

Good effort though, in any case.

Thank you :-)

Jacob

tmorrow · 2007-04-15, 11:56

Quote:

Originally Posted by Yxine

http://zip.larin.name/MasterPrime2006-Stats.zip (1.183.606 bytes)
http://zip.larin.name/MasterPrime2006-Statuses.zip (581.367 bytes)

These two files are no longer available at that location, producing file not found errors. Have they been relocated? Does anyone know where I can get them?

2007-01-04, 08:30	#1
Yxine "Aleksandr" Nov 2005 Russia 2×3²×5 Posts	Full report for 2006 year I generated the usual year report of done work: http://zip.larin.name/MasterPrime2006-Stats.zip (1.183.606 bytes) Also, here is a full statuses reports (113 reports for 11 years): http://zip.larin.name/MasterPrime2006-Statuses.zip (581.367 bytes)

2007-01-04, 15:28	#4
S485122 Sep 2006 Brussels, Belgium 2×3×281 Posts	On the Prime Net Top Producer page, one can get the full report Top Producers Awards, Full Length Report As of 2007-01-04 14:00 UTC the totals where Code: LL P90 Exponents Fact.P90 Exponents P90 CPU hrs/day CPU yrs LL Tested CPU yrs w/ Factor 1 958 357,026 888 875 60 071,380 103 786 15 202 356,250 !!! I use the ISO character for the decimal separator : the comma. (Lazily computed by totalling in Excel after breaking the report over two sheets.)

2007-01-05, 03:13	#5
jinydu Dec 2003 Hopefully Near M48 2×3×293 Posts	Thanks for the figures S485122. So GIMPS has indeed passed the 2 million CPU years mark. It would interesting to see how that figure has changed over the course of months and years. By the way, if all you did was add all the entries together, the P90 CPU hrs/day figure is incorrect. For instance, I have 3094.60 CPU hrs/day. This is my average speed since joining on Nov. 18, 2003; not my average speed since the start of GIMPS, which is obviously several times lower. So the figure of 15202356.250 CPU hrs/day is definitely an overestimate. Also, sorry for being nitpicky, but even the figures for LL P90 CPU years and Factoring P90 CPU years are not exactly correct because some GIMPS participants do not use Primenet. Still, I would think that it is a good approximation. Last fiddled with by jinydu on 2007-01-05 at 03:16

2007-01-05, 12:47	#8
S485122 Sep 2006 Brussels, Belgium 1686₁₀ Posts	GIMPS statistics on 2007-01-01 As promised the total from the GIMPS pages. Since the files on ftp://mersenne.org/gimps/ are dated from 2007-01-01 16:10 to 16:54 that must be the date of the GIMPS Top Producers statistics as well. Code: Who P-90 CPU years Exponents tested Total 1- 100 649 331,69 332 470 Total 101-10 000 1 224 169,41 688 741 Total 10 001-20 000 123 557,42 106 847 Total 20 001-30 000 42 975,92 51 234 Total 30 001-40 000 15 174,87 51 813 Total 40 001-50 565 3 475,56 36 656 Grand Total 2007-01-01 2 058 684,87 1 267 761 The difference with PrimeNet is big indeed. Code: Grand Total 1 958 357,026 888 875

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Year Over Year TF Progress	petrw1	Factoring	3	2013-03-20 19:34
Full exponent report	ahmerali	PrimeNet	3	2008-01-30 05:59
k*2^n-1 primes in 2006	Kosmaj	Riesel Prime Search	2	2007-01-02 04:31
Supercomputing 2006	ixfd64	Lounge	1	2006-11-30 23:34
2006	R. Gerbicz	Puzzles	45	2006-09-19 19:32

2007-01-04, 09:57	#2
jinydu Dec 2003 Hopefully Near M48 6DE₁₆ Posts	Here is a question I am curious about: What is the total amount of work done by GIMPS in its entire lifetime so far? The estimate given at http://mersenne.org/ips/stats.html is 11 months out of date.

2007-01-04, 15:08	#3
Yxine "Aleksandr" Nov 2005 Russia 2×3²×5 Posts	I can't count this... sorry!

2007-01-05, 13:11	#9
jinydu Dec 2003 Hopefully Near M48 2·3·293 Posts	How did you determine the number of CPU years completed? If I am not mistaken, only the number of CPU years remaining is shown. Also, what about the factoring CPU years completed. Good effort though, in any case.