The enormous size of the diagonal Latin squares space makes it unfeasible to enumerate all its objects straightforwardly in reasonable time. So, in order to discover the structure of this space, sophisticated search methods are needed. In RakeSearch project, we implement an application that picks up separate pairs of mutually orthogonal DLSs, which allows to reconstruct full graphs of their orthogonality. You can contribute to our research by running a free program on your computer.
For account creation please use a Crunch_4Science invitation code.
New version of application for Rank 10
We deploy a new version of the application for search in rank 10 space. It will be faster than the previous version due to the earlier exclusion of some combinations that do not lead to a result - we add a filter for excluding combinations with the same rows in one position. Example of a similar combination you can view in July 2019 totals.
We expect that the results for the same workunit processed by the different versions of the application may differ and marked as incorrect. But in our estimation, the number of such results will be relatively small - about several tens, maybe 100 or 200. If you face with such workunits - post these numbers in this thread, please!
Thank you for attention for the project and participation!
7 Aug 2019, 21:08:17 UTC · Discuss
July 2019 totals
As you know in July we started a new search - permutational diagonal Latin squares (ODLS) of rank 10. Due to lack of information about the existence of these pairs, we include into results data about found partially orthogonal pairs.
No ODLS of rank 10 found yet, but we collected the following statistics about partially orthogonal pairs from 15318 workunits processed in July :
Degree 81: 508008
Degree 82: 121883
Degree 83: 23545
Degree 84: 4979
Degree 85: 684
Degree 86: 163
Degree 87: 10
Degree 88: 4
Degree 89: 0
Degree 100: 0
One of the pairs with degree 88 looks some funny:
but the other - more usual:
We are still thinking about badges for new search (except "credit badges" of course) but for this month totals we want to congratulate the participants, which computers found pairs with degree of orthogonality of rank 86 or higher:
terry and [AF>France>TDM>Picardie>Oise] docmic92 (L'Alliance Francophone),
arkiss and [AF>Amis des Lapins] Phil1966 (L'Alliance Francophone),
Heikki and [AF>France>Savoie] Tutu1986 (L'Alliance Francophone),
raddoc and [AF>WildWildWest]Sebastien (L'Alliance Francophone),
Progger and [AF>Libristes]cguillem (L'Alliance Francophone),
zgf2022 and Michael H.W. Weber (Rechenkraft.net),
Boliver and zelandonii (L'Alliance Francophone),
Leonmer and LeChat51X (L'Alliance Francophone),
Solo Man and Steve Dodd (The Planetary Society),
dabarlow and Rhodan71 (L'Alliance Francophone),
DRandolph and eeeeee (Universite de Sherbrooke),
yippee0226 and dpc_kluizenaar (Dutch Power Cows),
GESUCHTMAN and arjanscholl (Dutch Power Cows),
faliqisbah and BlisteringSheep (Ars Technica),
Mark Sechin and Fateev Aleksey (SETIKAH@KOREA),
AlexxSaigon and [SG-2W]Kurzer (SETI.Germany),
Kevin (USA) and killik (Czech National Team),
Administrator and DoctorNow (BOINC@Heidelberg),
[B^S]ABoskovich and Shaman (Czech National Team),
FourOh (SETI.USA) and Ralfy (BOINC Confederation),
Casey ([H]ardOCP) and respawner (Rechenkraft.net),
saiz (Team China) and [B@P] Daniel (BOINC@Poland),
STE\/E (BlackOps) and Rudy Toody (TeAm AnandTech),
Buga1 (Calm Chaos) and BritishBob (Overclock.net),
NeoGen (AMD Users) and ReaDy (Astronomy.Ru Forum),
fix (Crystal Dream) and [SG]steini (SETI.Germany),
cc-smokey (CC-Team) and Vincent Dark (Team China),
AGG (Crystal Dream) and hoarfrost (Crystal Dream),
Werinbert (Free-DC) and Roadranner (SETI.Germany),
morgan (Team Norway) and Landjunge (Planet 3DNow!),
axels (SETI.Germany) and Paulteo (Crunching@EVGA),
Derivator (Gridcoin) and CoolAtchOk (Russia Team),
inugamikoubou@hangul and muzzdiez (SETIKAH@KOREA),
VogeL (SETI.Germany) and Spritex (BOINC@Denmark),
jozef j (Russia Team) and Kodeks IT (TSC! Russia),
citerra (Russia Team) and Dario666 (BOINC@Poland),
lydy1993 (Team China) and [B^S] fred (BOINCstats),
JagDoc (Planet 3DNow!) and kim-b-h (BOINC@Denmark),
Timothy Mullican (USA) and Dingo (BOINC@AUSTRALIA),
Gipsel (Planet 3DNow!) and Peter (BOINC@AUSTRALIA),
Soderstromm (Gridcoin) and sergamena (Russia Team),
SerGiani (Russia Team) and whizbang (Ars Technica),
Novalor (SETI.Germany) and RFGuy_KCCO ([H]ardOCP).
(Their number is 86 also!)
Thank you for attention to the project and participation!
3 Aug 2019, 8:27:01 UTC · Discuss
New run for RakeSearch R10
We created the first bunch of tasks for new search in rank 10 space. Currently - under Linux x86-64. Renewed search use a new application, runtime increased in several times and new application return more data than previous. If tasks processing is successful tomorrow we add details! :)
Great thanks for Daniel from BOINC@Poland for attention, scientific critical perception and help!
15 Jul 2019, 19:49:07 UTC · Discuss
Project results for rank 9
Today, on July 11, our project turns 23 months! In June, the latest search results were obtained on the DLS space of rank 9 and we begin to sum it up! The first and very important one is the number of diagonal Latin squares of rank 9. Each result recorded the number of processed DLSs and now, summing them up, we got the number 5 056 994 653 507 584, which coincided with the result of our colleagues from Gerasim@Home, listed in OEIS. Since the computational modules were developed independently, they did not use the same algorithms, the division of the space into tasks was different, and the optimized Daniel’s application was also useds, this is a very important result confirming the first assessment and the fact that the search covered all DLSs of rank 9!
Another result is that a paper on the project has been accepted to a “Communications in Computer and Information Science” series as one of selected papers of a conference “Russian Supercomputing Days”. It will be published in 2019 and we will share the link with you.
Discussion in Russian is here.
11 Jul 2019, 4:50:45 UTC · Discuss
R10 search temporary stopped!
We stopped the sending of tasks for R10 RakeSearch due signal from Daniel about incorrect fragment of code of new application. Now we investigate the problem and after this - (presumably) will build a new application.
25 Jun 2019, 12:53:05 UTC · Discuss
©2019 The searchers team