Strange credit calculation for Spectra of Latin squares

Message boards : Number crunching : Strange credit calculation for Spectra of Latin squares
Message board moderation

To post messages, you must log in.

AuthorMessage
Antoshka
New member

Send message
Joined: 12 May 25
Posts: 3
Credit: 11,942
RAC: 817
Message 2000 - Posted: 18 May 2025, 23:08:49 UTC

It looks like credit = runtime / 432.
This means that faster machine will get less credit than slower machine for the same amount of work.
Even more, the run time is wrong. It counts run time from last checkpoint only. In reality first 8 tasks on my screenshot ran for about 56 hours, but this computer wasn't powered on continuosly.
ID: 2000 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
hoarfrost
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Help desk expert

Send message
Joined: 11 Aug 17
Posts: 669
Credit: 24,912,994
RAC: 16,133
Message 2002 - Posted: 19 May 2025, 9:55:47 UTC

Hi! When granted credit is proportional to amount of calculations, for each computer granted credit = run_time / k. But k are different for different computers. :)
ID: 2002 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : Strange credit calculation for Spectra of Latin squares

©2025 The searchers team, Karelian Research Center of the Russian Academy of Sciences