Computing analysis for lichess players requires a lot of CPU power, and we're running short of it.
If you want to help lichess and have a spare server, now is the time to contribute to the lichess adventure! I will take care of installing lichess AI on your server, and you will be mentioned on the
lichess.org/thanks page.
So, who wants to donate CPU time to free chess?
side question : is the number of analysis we can do each day limited ?
No. Everybody can do as many analysis as they want, but only one at a time.
Thibault, would this work just by using idle CPU cycles (sort of like Folding@home) or be able to be turned on or off (like the Stockfish testing framework)?
I ask because I don't have spare, unused computers, but I certainly have some unused CPU time I could donate if that's consistent with how it works.
Thanks again for all you do!
The idea of Fishtest pleases me well coupled with a link "Get involved" as for
Stockfishchess.orgAnd as I've said in a previous post, a partial analysis is often sufficient (Analysis stops at the first blunder)
I hope that you will find a solution.
Regards
I guess my raspberry pi won't be worth the trouble...
A dristributed solution is far too complicated to setup.
Thanks pbnn but that's not what pis are for, indeed ;)
May be some distributed computing as bitcoin?
For example, everything who want install AI on machine and computing some games which are sended from server?
Unfortunately I don't have a server you can use. What's the plan for scaling the site, though? You'll probably have to keep adding servers as the active userbase grows, so even if you can find someone willing to let you use theirs, you will probably be back in this situation again before long.
Ah, it's a shame the distributed model is too difficult to implement, although it's understandable, of course.
What are the specifications for the current hardware backing the analysis?
I'm trying to get an idea of what sort of headroom I could provide with different CPUs if I go shopping.
Thanks again!
