I added PADLS application on my BOINC server. https://boinc.tbrada.eu/

It is open to everybody, mainly because I do not know how to not open it for everybody. Anyway, you can join and perform some computation.

The PADLS application is running Pseudo Associative Diagonal Latin Squares experiment by Natalia Makarova. Application is compiled for the following systems:

Linux on amd64 CPU and armv7l. This application requires Boinc v7 LibC v6 and ncursesw v6.

Windows 64-bit.

I loaded 1000 WUs from ms. Makarova. These were loaded with quorum requirement of 2, because I am not fully confident with the application. Might not have been the best idea to load all 1000 at once with quorum 2. We will see how it goes and if it goes for long, I will at how to reduce the quorum to 1.

There are forums, but It would be better to stay with discussion in this thread.

There is no validation yet, so completed results will just sit in the DB. But that is not a problem, they will be validated in batch and credit assigned. The FLO estimate is wrong at first, but it will improve.

ms. Makarova: please post short description or link of the experiment.

---

ms. Makarova: please post short description or link of the experiment.

Tomas Brada
I present a short description of the PADLS experiment.
If this is not enough, tell me.
You can edit the description.

-------------
Pseudo-associative DLS of order 10 is not a fully associative DLS - with some violation of associativity.
Associative DLS of order 10 does not exist.

The class of pseudo-associative DLSs of order 10 is very large.
As a result of the experiment, it was found that pseudo-associative DLSs give many orthogonal DLSs when processed by their special program by Belyshev.

My first generator gave such a pseudo-associative DLS of order 10



See
https://boinc.progger.info/odlk/forum_thread.php?id=51&postid=1247#1247

In an experiment with pseudo-associative DLS,
for example, a unique solution was found, noted in OEIS
https://oeis.org/A287695

At the moment two generators of pseudo-associative DLS are working in the experiment.

To increase the speed of processing, more computers should join the server.
At the moment, my 3 computers and a unknown user Vasiliy is computing. And Vasiliy is sending only failures.
It would help if someone could look at the psevdoass.exe from here and find which DLL is it missing.

Tomas Brada
do not hurry.
Over time, users will be. Your project is still little known.

I wrote an invitation to ice00, but he is not participating yet.

New Windows app build, should fix the dll problem.

I am quite curious why there are two generators?

Already 796 new cf odls. Preliminary results:
https://boinc.tbrada.eu/download/pre_results_10_0.txt
https://boinc.tbrada.eu/download/prep_results_10_0.txt
count[1]: 779
count[2]: 17
All: 796
CF co-squares: 789

I am quite curious why there are two generators?

Generators generate pseudo-associative DLS with varying degrees of associativity violation.

In the illustration above, you see the DLS from the first generator.
In such DLS, there is a violation of associativity in only two central rows.
There are quite a few DLS.

I will show illustrations from the second generator tomorrow.
In DLS from the second generator, the violation of associativity in the four central rows.

Already 796 new cf odls. Preliminary results:
https://boinc.tbrada.eu/download/pre_results_10_0.txt
https://boinc.tbrada.eu/download/prep_results_10_0.txt
count[1]: 779
count[2]: 17
All: 796
CF co-squares: 789

How many WU have been processed?

In this experiment, a large repeatability of results.
Do not forget that the experiment is performed for the second year in my manual project.
The main part of the solutions is in the database of manual project.
Tomorrow I will publish a database of manual project (with the exception of the part containing 67879 CFs ODLS, this part has already been published).

Tomas Brada
at the moment you should check the uniqueness of the solutions relatively this part of the database

10. Addition to the database from the manual project

67879 unique CF ODLS

May 1 - November 4, 2018

©2018 (C) Natalia Makarova & Demis & ice00
©2018 (C) Progger & Stefano Tognon (ice00)

Link
https://cloud.mail.ru/public/EAbb/fmt9Pg1RV

Created on November 5, 2018

This part is included in the total database you have.
If possible, check the uniqueness of the relative total database.

I checked the solutions for uniqueness relative of database of the manual project.
Found 369 unique CFs ODLS.

I will show illustrations from the second generator tomorrow.
In DLS from the second generator, the violation of associativity in the four central rows.

This is a pseudo-associative DLS of order 10 from the second generator



Possible and the third generator!
Do you guess? :)

Addition to the database from the autonomous subproject

21203 unique CFs ODLS

November 5, 2018 - February 10, 2019

©2019 (C) Natalia Makarova & participants of autonomous subproject

Link
https://cloud.mail.ru/public/6LMT/Ngt15i43n

Created on February 10, 2019

Tomas Brada
please add this portion of solutions to the common database.
I checked the uniqueness of these solutions only within the framework of the ODLK project and the manual project.

Row "6 1 9 3 2 7 5 4" gave 245 cf odls. Usually they give around 20.

XAVER has from 400 WUs:

minimum - 1 CF
maximum - 122 CFs.

WU without solutions no.

Results will be processes once a batch finishes.

Row "6 1 9 3 2 7 5 4" gave 245 cf odls.

I processed this WU.

Найдено марьяжных КФ:
count[1] = 1027
count[2] = 514
count[4] = 19
Всего: 1560
Найдено соквадратов: 2131
КФ соквадратов: 1560

There are several “symmetrical” solutions in this portion.

Поиск чистых симметрий

Введено ЛК: 1560
Введите код симметрии:
all
Квадратов с симметрией (1,31,31) найдено: 6 они записаны в файл symm_1_31_31.txt
Квадратов с симметрией (4,31,31) найдено: 3 они записаны в файл symm_4_31_31.txt
Квадратов с симметрией (16,16,16) найдено: 1 они записаны в файл symm_16_16_16.txt
Квадратов с симметрией (16,31,31) найдено: 1 они записаны в файл symm_16_31_31.txt

Время поиска: 1.825 сек

Unfortunately, almost all solutions are repeated, only 13 unique CFs ODLS.

However, the result is very interesting, it confirms that pseudo-associative DLSs give a lot of ODLS, including groups of pairs of ODLS.

Tomas Brada
would not a quorum reset result in lost solutions?

For example, the user XAVER has the following WU

2878 3000 11 Feb 2019, 10:57:21 UTC 11 Feb 2019, 11:00:58 UTC Завершён, ожидает проверки 2.18 0.03 ожидание Pseudo Associative DLS v1.04 windows_x86_64

This WU has been processed and is awaiting checking.

Tomas Brada
I do not understand why new users are not joining the project.
Is the project open to new users?

Maybe you want to perform an experiment with only three participants?

Tomáš Brada
Всего очков 40,437
XAVER
Всего очков 7,060
valap
Всего очков 1,443

To attract more users, you should write about the project. For example on boinc new projects subforum.
https://boinc.berkeley.edu/forum_forum.php?id=11
Also facebook/twitter/vk, http://boincstats.com