OBD levels progress history

[kriesel]

Thanks to numerous posts by various folks in multiple threads here, I've pieced together a partial history of when the various OBD levels were reached. And of course there is a spreadsheet with a chart, fit, references, etc. There are some substantial gaps. Please comment here or by PM with any potential gap-filling info or links or corrections.

[Uncwilly]

As of Oct 26, 2011 OBD was at Level 17.02
It was at 3.3 on April 26, 2004
Do you want some finer data? And have you tried the Wayback Machine?
I have some data from: 2/11/2005, 1/12/2005, 12/10/2004, 11/12/2004, 10/15/2004, 9/17/2004, 8/13/2004, 7/16/2004

Level 17	Eighteen Candidates to 84 bits	In Progress
Level 17	Seventeen Candidates to 83 bits	 22 July 2011
Level 16	Sixteen Candidates to 82 bits	4 July 2011
Level 15	Fifteen Candidates to 81 bits	26 Nov 2010
Level 14	Fourteen Candidates to 80 bits	15 June 2010
Level 13	Thirteen Candidates to 79 bits	6 June 2010
Level 12	Twelve Candidates to 78 bits	16 Sept 2009
Level 11	Eleven Candidates to 77 bits	28 May 2009
Level 10	Ten Candidates to 76 bits	16 May 2009
Level 9	Nine Candidates to 75 bits	3 Dec 2007
Level 8	Eight Candidates to 74 bits	10 May 2006
Level 7	Seven Candidates to 73 bits	28 Jan 2005
Level 6	Six Candidates to 72 bits	23 July 2004
Level 5	Five Candidates to 71 bits	22 May 2004
Level 4	Four Candidates to 70 bits	03 May 2004
Level 3	Three Candidates to 69 bits 12 Apr 2004
Level 2	Two Candidates to 68 bits	07 Apr 2004
Level 1	One Candidate to 67 bits	26 Mar 2004

[Uncwilly]

I checked, the Wayback machine has some better data:
https://web.archive.org/web/20120215...h/Billion.html
The Italian site did not have a good status page that I could see for very long.

[kriesel]

Thanks for the wayback machine suggestion.
wblipp's eleven-smooth billion page is a dead link
home.earthlink.net/~elevensmooth/Billion.html wayback coverage 2004-2019
so is ET_'s http://www.gimps.it/billion/billion.htm wayback has 2004 and one in 2021 is a 404 error
and http://www.moregimps.it/billion/billion.htm; see wayback.org 2004-2010
so is the 100M page http://www.geocities.com/onehundredmdpp/index.htm
up but outdated: http://2721.hddkillers.com/graph/ wayback sometimes does not show the graph, only the abscissa values and a little other text, or it's flipped and the popup text gets hidden;
https://www.mersenne.ca/obd is the status page now; wayback has several snaps 2020-2021
I went through a lot of those (all sites above).

Not looking for fine detail within levels or a data dump. Just trying to pin down integer level history/milestone dates, and a little intermediate progress where levels have spanned years (2014-2019?!).

[pinhodecarlos]

Why the model fitting?

[kriesel]

Why not? The portions where level declines with increasing time are silly, but the whole subproject is somewhat.

[lavalamp]

Originally Posted by kriesel

Thanks for the wayback machine suggestion.
wblipp's eleven-smooth billion page is a dead link

For a little while I mirrored and updated the original page here:
https://2721.hddkillers.com/OBDElevenSmooth/

The last update I made to it was 13 April 2020. This page is the source my graph and table pull from, hence why they are more than slightly out of date currently.

[kriesel]

Reached level 21.21 on 2021-10-21. Level 22+ is expected in a few days.

[kriesel]

level 21.22 -> level 22.03 2021-10-27

[kriesel]

Level 22 has been passed.
If, with no factors found for the counted survivors,
level 23 is 23 candidates TF to 89 bits,
level 24 is 24 candidates TF to 90 bits,
level 25 is 25 candidates TF to 91 bits,
level 26 is 26 candidates TF to 92 bits,
and OBD TF factoring depth limit is 92 bits, (as is mfakto's),
then,
is level 27, 27 candidates P-1 stage 1 complete with sufficient B1?
Level 28, 28 candidates P-1 stage 2 also complete with sufficient B1 & B2?
Level 29 (someday, one can dream), 29 candidates PRP/GEC/proof-generation completed?
Level 30 (also requiring new software functionality), 30 candidates PRP proof verified? (Or alternately an entire very costly full length PRP DC for each, ugh.)

That would imply carrying through on ~33 current candidates, since one or two (~33 x (92-88) / 90 ~1.5) will likely get factored in TF, and ~4% x 33 ~1.3 in P-1. So, all in range 3321928171 to 3321929987.

[clowns789]

The problem with levels 27-30 that you outlined is that it encourages the two P-1 stages to be separate and thus many of the second stages will likely be done on different systems than the first stages. I believe after Level 26 the project should split into a "wide" search for those who want smaller GPU work units to factor more exponents to lower bit depths, and a "deep" search with P-1 and PRP.

I think P-1 stages could be assigned separately, although it shouldn't be encouraged by the way the level system is structured; in other words, it shouldn't encourage participants to run 27 first stages before doing any second stages.