PADLS Total subproject info

Message boards : Information : PADLS Total subproject info
Message board moderation

To post messages, you must log in.

AuthorMessage
Tomáš Brada
Project administrator
Volunteer developer
Avatar

Send message
Joined: 3 Feb 19
Posts: 632
Credit: 445,023
RAC: 12
Message 3650 - Posted: 14 Oct 2019, 20:28:44 UTC
Last modified: 19 Sep 2020, 11:59:35 UTC

Sub-project PADLS Total.

The PADLS Total subproject aims to find new pseudo-associative DLS as described by Natalia Makarova in this thread. Not a groundbreaking amount, but there are good results produced every day. Moreover this subproject produces a seemingly endless stream of workunits for us.

Help

If you have any issue, do not hesitate to create a new thread in Help and new Issues section. You can also continue discussion in any related thread. But if you are not sure, start a new thread, it is easier for me to deal with.

Application

Programming talk and development updates are in this thread. The application is built upon code by Belyshev.

The workunit length is variable. I can make them longer or shorter in computation time. By default it would run very very long time, but every time it checkpoints, it checks whether it already done enough work and if so, it finishes. Same when you shut down boinc.
You can set checkpoint interval in your BOINC Manager. By default it is set to one minute, but if you set it to one hour, the workunit will run up to one hour longer than designed. This is OK for the project: the results are fine and credit is assigned adequate to work performed.

In case your task is nearing or already passed a deadline, it finishes before reaching it's target. This is to prevent boinc aborting it and wasting all computation. This may confuse the boinc client and you, but you'll still get credit.

The run time is dominated by integer calculations and branches. No floating point arithmetic and little vectorization. This means CPU running this task will run cooler compared to cpu running for example LLR.

Credits

Credit is assigned by validator/assimilator program based on the real work performed by the application.

The app is counting number of three key steps in the computation. Each of those was benchmarked on my computer, from this benchmark times, a set of relative weights were assigned to the counters. Lastly the result was rescaled to cobblestone units according to Berkeley recommendation of 200 credits per GFlop. More info can be found in this heated thread. Thanks Michael Goetz for finding the bug.

Server

The search space has been split into so called segments. From each segment one workunit is generated at first. The workunits can be made any size, but are currently set to run approximately two hours on my computer. Once result of a workunit is returned to the server, it is picked up by the validator program. This program checks the result validity, saves the result into the database and generates a new workunit to continue processing the segment.

This is batch 15, 21 and 22 on the server_status page.
Batch 21 (tot5-i) is for initial workunits - it consists of workunits in segments that have not yet been computed.
Batch 22 (tot5-m) is for main workunits - those are continuation of previous result in a segment.

Results

Results are exported twice a day, and they are available for download as text or sql.
The text file tot_odlk_plain.txt contains only results after database id 526126 (november 2019).
Earlier are stored in a static file tot_odlk_p201911.txt.

The search algorithm is not the most efficient, but gives reasonable results.

Source Code.
ID: 3650 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Natalia Makarova
Project scientist
Avatar

Send message
Joined: 8 Feb 19
Posts: 356
Credit: 0
RAC: 0
Message 4341 - Posted: 2 Dec 2020, 2:27:51 UTC - in response to Message 3650.  
Last modified: 2 Dec 2020, 2:39:14 UTC

Sub-project PADLS Total.
. . . . .

Source Code.

The link is not valid.

Is the PADLS TOTAL rule 51 subproject dead?
I cry like a funeral. :(
ID: 4341 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Natalia Makarova
Project scientist
Avatar

Send message
Joined: 8 Feb 19
Posts: 356
Credit: 0
RAC: 0
Message 4343 - Posted: 2 Dec 2020, 2:53:50 UTC

Tomáš Brada
My assistant and I wanted to try to export the PADLS TOTAL rule 51 subproject to the yoyo@home platform.
But you deleted the sources.
It is now impossible to export.
It is sad.
ID: 4343 · Rating: 0 · rate: Rate + / Rate - Report as offensive

Message boards : Information : PADLS Total subproject info

©2021 Tomáš Brada