You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A bit more complex than the "bind/locking" request: The ability to create a pool of agents and moving agents in and out of the pools, whereas each pool works on a given task from the priority list. This would work out as follows:
a) all agents join a pool upon creation
b) you can create a new pool and move agents between existing pools
c) each pool works on a different task, making it possible to easily adjust the number of agents / tasks and parallel tasks in the network
This would require some changes in the "request new chunk" code and one of the problems to take care of would be to prevent a subset in a pool from idling when there are no more chunks in the pool's task, e.g., only two tasks { I, II }, agent pools {1, 3, 4} and {2, 5, 6, 7, 8, 9} and agents {5, 7, 9} finish their chunk and the only chunks left belong to the task the other pool works on.
Apart from that, I think that pooling would be a great feature :)
The text was updated successfully, but these errors were encountered:
i understand the idea and it is a good one, however i think that having so many agents to need this is a very niche market for hashtopus users. i would also have trouble properly testing it since i only have 2 machines available.
A bit more complex than the "bind/locking" request: The ability to create a pool of agents and moving agents in and out of the pools, whereas each pool works on a given task from the priority list. This would work out as follows:
a) all agents join a pool upon creation
b) you can create a new pool and move agents between existing pools
c) each pool works on a different task, making it possible to easily adjust the number of agents / tasks and parallel tasks in the network
This would require some changes in the "request new chunk" code and one of the problems to take care of would be to prevent a subset in a pool from idling when there are no more chunks in the pool's task, e.g., only two tasks { I, II }, agent pools {1, 3, 4} and {2, 5, 6, 7, 8, 9} and agents {5, 7, 9} finish their chunk and the only chunks left belong to the task the other pool works on.
Apart from that, I think that pooling would be a great feature :)
The text was updated successfully, but these errors were encountered: