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
Ability to partition workload seamlessly over multiple processes (uses all cores), if applicable to partitionning over nodes instead of cores would allow parallel computing on the edge.
GenStage
Producer - producer/consumer - consumer schemes based on gen_server wheee a consumer asks for specific amount of tasks, and the producer cannot overflow it. If multiple subscribers exist for a single producer, the fastest ones will sustain the heavier loads based on observed performance.
During the consumption, the consumers ask incrementally for more tasks from producers after some portion e.g. 50% of tasks are completed s.t. producers are never idle and keep sending tasks at adapted pace, hence there is continuous bidirectional communication and no waiting on any side.
The text was updated successfully, but these errors were encountered:
Flow
Ability to partition workload seamlessly over multiple processes (uses all cores), if applicable to partitionning over nodes instead of cores would allow parallel computing on the edge.
GenStage
Producer - producer/consumer - consumer schemes based on gen_server wheee a consumer asks for specific amount of tasks, and the producer cannot overflow it. If multiple subscribers exist for a single producer, the fastest ones will sustain the heavier loads based on observed performance.
During the consumption, the consumers ask incrementally for more tasks from producers after some portion e.g. 50% of tasks are completed s.t. producers are never idle and keep sending tasks at adapted pace, hence there is continuous bidirectional communication and no waiting on any side.
The text was updated successfully, but these errors were encountered: