|
@@ -1274,6 +1274,7 @@ TODO: improve!
|
|
|
|
|
|
|
|
@menu
|
|
@menu
|
|
|
* Data management::
|
|
* Data management::
|
|
|
|
|
+* Task submission::
|
|
|
* Task scheduling policy::
|
|
* Task scheduling policy::
|
|
|
* Task distribution vs Data transfer::
|
|
* Task distribution vs Data transfer::
|
|
|
* Power-based scheduling::
|
|
* Power-based scheduling::
|
|
@@ -1314,6 +1315,15 @@ OpenCL to make the allocation itself and pin the corresponding allocated
|
|
|
memory. This is needed to permit asynchronous data transfer, i.e. permit data
|
|
memory. This is needed to permit asynchronous data transfer, i.e. permit data
|
|
|
transfer to overlap with computations.
|
|
transfer to overlap with computations.
|
|
|
|
|
|
|
|
|
|
+@node Task submission
|
|
|
|
|
+@section Task submission
|
|
|
|
|
+
|
|
|
|
|
+To let StarPU make online optimizations, tasks should be submitted
|
|
|
|
|
+asynchronously as much as possible. Ideally, all the tasks should be
|
|
|
|
|
+submitted, and a mere @code{starpu_task_wait_for_all} call be done to wait for
|
|
|
|
|
+termination. StarPU will then be able to rework the whole schedule, overlap
|
|
|
|
|
+computation with communication, manage accelerator local memory usage, etc.
|
|
|
|
|
+
|
|
|
@node Task scheduling policy
|
|
@node Task scheduling policy
|
|
|
@section Task scheduling policy
|
|
@section Task scheduling policy
|
|
|
|
|
|
