There is Job size matching feature of boinc. With variable-length workunits, we might be able to use it in the future.

The current maximum CF DLS in rule 51: T23CdsnYRJXYZAMm1RAJi2RE2

As far as I understand, the CFs you are testing do not go in order.
This is a spread over some large area of the ruler.

No, they are not tested in order in the grand scheme. Within a workunit they go in order, but those workunit start at points that are spread out.
The CF i posted might not actually be the maximum found, because I made a mistake in the query from the database.

I added 3 segments to the database for rule 15 and for 38. Those are just a tiny fraction (0,0016%) of all in those rules and are spread around. It is just for testing, but maybe, they give interesting results.

It's good.

Is it possible to run separate Applications for rules 15 and 38?
Can the server handle the load?

Later, we should talk to RakeSearch project, because they are too studying LK10.

Later, we should talk to RakeSearch project, because they are too studying LK10.

Sorry, I did not understand.
Why do we have to talk about something with the project RakeSearch?
We have nothing to do with this project.

If they want to run an ODLS of order 10 search, that's their business.
It can not disturb us!

I found it

Будущее проекта RakeSearch (Ru)
Дорогие участники!

Поиск ОДЛК 9-го порядка на 99.9% близок к завершению. Запущены тестовые задания для 10-го порядка, пока — только для платформ Linux x86-64 и Windows x86-64. В среднем одно такое задание осуществляет перебор большего числа квадратов за то же самое время. Пока что ни в одном задании не удалось обнаружить пары перестановочных квадратов 10-го порядка. Мы планируем провести тестовый поиск по небольшой части пространства квадратов (возможно, порядка миллиона), чтобы понять перспективы. Если таких пар не найдется, то проект разумно будет завершить, чтобы не занимать ваши процессорные мощности зря. Если найдутся — это будет новым результатом и базой для нового поиска.

В любом случае, проект не дублирует расчеты Gerasim@home и ODLK/ODLK1 для 10-го порядка.

See
https://rake.boincfast.ru/rakesearch/

I draw your attention to this

В любом случае, проект не дублирует расчеты Gerasim@home и ODLK/ODLK1 для 10-го порядка.

I am not worried about duplication. I am just interested in what this "Rake" search is about. Rake is a gardening tool.

With the increased deadline, crunchers cache much more tasks now. There are 31146 tasks in progress and the ready to send is getting down. I can easily generate more when it gets to it.

I am just interested in what this "Rake" search is about. Rake is a gardening tool.

It was “Rake” for ODLS of order 9.
I do not see the point in "Rake". It would be quite possible to compile a complete database of CFs ODLS of order 9.

Now they use “Rake” for ODLS of order 10.
I guess it will be empty here.
There are “permutation pairs” of OLS of order 10.
For example,

 0 1 2 3 4 5 6 7 8 9
 1 0 5 8 3 2 7 9 6 4
 2 5 0 6 9 4 3 8 1 7
 3 8 6 0 7 1 5 4 9 2
 4 3 9 7 0 8 2 6 5 1
 5 2 4 1 8 0 9 3 7 6
 6 7 3 5 2 9 0 1 4 8
 7 9 8 4 6 3 1 0 2 5
 8 6 1 9 5 7 4 2 0 3
 9 4 7 2 1 6 8 5 3 0

 0 1 2 3 4 5 6 7 8 9
 2 5 0 6 9 4 3 8 1 7
 3 8 6 0 7 1 5 4 9 2
 4 3 9 7 0 8 2 6 5 1
 5 2 4 1 8 0 9 3 7 6
 6 7 3 5 2 9 0 1 4 8
 7 9 8 4 6 3 1 0 2 5
 8 6 1 9 5 7 4 2 0 3
 9 4 7 2 1 6 8 5 3 0
 1 0 5 8 3 2 7 9 6 4

“Permutation pairs” of ODLS of order 10 I have not met.

By the way, a good task: check for the presence of “permutation pairs” of ODLS of order 10 in the existing database.

Tasks ready to send 0
Задания обрабатываются 33417

Tomáš Brada
Why do tasks end from time to time?

There were more tasks in progress than there was open segments. I thought that was enough tasks, but it was not. I opened 24000 more segements, so there should be enough tasks now.

I opened 24000 more segements, so there should be enough tasks now.

Thanks.

Interesting oddity in MYSQL:
There is index: ind_res_st on columns (server_state, priority)
and second index: res_app_state on (appid, server_state, outcome, validate_state)
The query: select * from result where appid=7 and server_state=5 and outcome=1 and validate_state=0 limit 250
will most often use the ind_res_st, which is slow (184s), instead of the index that matches the query exactly (0.1847s).
This is crazy!

Why not publish the source code?
Some boinc projects have benefits from volunteer developers.

The source code is public: github

The source code is public: github

Great!
(maybe insert a link in home page)

A limit to server bandwidth has been set up.

How many CFs are there in rule 51?

Belyshev has published the counter_zapoln program.
https://boinc.ru/forum/topic/issledovaniya-svojstv-latinskih-kvadratov/#postid-355

An example of the program for three rows of rule 51

Номер линейки: 51
Число ячеек  : 24
############################################################
3071716816############################################################
6131181164############################################################
9215698675############################################################
12331561496############################################################
15453853711############################################################
18574690508##################
Вариантов    : 19550410058
Время поиска : 1896.87 сек
Продолжить? (Y/N):

Tomáš Brada
Can you rate the total CFs SNDLS in rule 51?

See the topic for more details
https://boinc.progger.info/odlk/forum_thread.php?id=130

22 tot5-m 93415 6108 12 9650 371

https://boinc.tbrada.eu/server_status.php

What do many errors mean?