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
We have the Par framework for data parallelism, but it would be nice to have something for basic task parallelism too (think GCD, futures, probably many models to choose from).
In its simplest form this would be a couple of data types Task and TaskGroup maybe, with the ability to receive callbacks when something is finished, or to wait for something to finish (or to wait for all/any of a group of things to finish). Marshaling of parameters and results probably.
There would probably have to be a notion of a worker that is suitable for a task (for task distribution and load balancing).
There might be a notion of canceling a job, though I'm not sure how to implement that, the job itself would have to poll to see if it's canceled, perhaps.
The text was updated successfully, but these errors were encountered:
We have the Par framework for data parallelism, but it would be nice to have something for basic task parallelism too (think GCD, futures, probably many models to choose from).
In its simplest form this would be a couple of data types Task and TaskGroup maybe, with the ability to receive callbacks when something is finished, or to wait for something to finish (or to wait for all/any of a group of things to finish). Marshaling of parameters and results probably.
There would probably have to be a notion of a worker that is suitable for a task (for task distribution and load balancing).
There might be a notion of canceling a job, though I'm not sure how to implement that, the job itself would have to poll to see if it's canceled, perhaps.
The text was updated successfully, but these errors were encountered: