Back to results
US 11,635,997 B2Granted
Dataflow optimization apparatus and method for low-power operation of multicore systems
AssigneeAjou University Industry-Academic Cooperation Foundation
InventorsHoeseok Yang, Hyeonseok Jung
Life of this patent
Estimating the term…
Abstract
The present disclosure relates to a dataflow optimization method for low-power operation of a multicore system, the dataflow optimization method including: a step (a) of creating an FSM including a plurality of system states in consideration of dynamic factors that trigger a transition in system states for original dataflow; and a step (b) of optimizing the original dataflow through optimization of the created FSM.
Claims8
A dataflow optimization method for low-power operation of a multicore system, the dataflow optimization method comprising steps of:
- creating a Finite State Machine (FSM) including a plurality of system states in consideration of dynamic factors that trigger a transition between system states for an original dataflow composed of a plurality of tasks, the dynamic factors indicating events that trigger the transition between the system states, wherein the dynamic factors are factors related to at least one of a Quality of Service (QoS) change and a power budget change, and wherein the plurality of system states are sequentially arranged in accordance with a QoS level or a power budget level;
- optimizing the original dataflow through optimization of the created FSM and storing the optimized original dataflow in a database;
- and
- in response to detecting a change in system state during runtime, controlling an unmanned flying object to autonomically adapt by extracting a response to the change in system state on the basis of the optimized original dataflow stored in the database, and providing the response to the unmanned flying object,
- wherein the step of creating further comprises creating the FSM such that a state transition is possible only between adjacent system states among the plurality of system states that were sequentially arranged,
- wherein the step of optimizing further comprises:
- mapping each of the plurality of tasks corresponding to the created FSM to at least one core of multiple cores to decide candidate mapping information;
- and
- determining whether predetermined conditions are satisfied by applying a priority-based list-scheduling to the plurality of tasks mapped to the multiple cores, and deciding the candidate mapping information as a final mapping information when the predetermined conditions are satisfied as a result of the determination,
- wherein in the step of mapping, each task of the plurality of tasks is designated with at least one slot among slots of the multiple cores as many as a maximum parallelism degree of the each task of the plurality of tasks, and
- wherein the predetermined conditions include a condition that sets latency to be a predetermined time or less in parallelizing the plurality of tasks scheduled through the multiple cores.
8 more elements
A dataflow optimization apparatus for low-power operation of a multicore system, the dataflow optimization apparatus comprising:
- a processor;
- and
- a memory storing instructions executable by the processor,
- wherein the processor is configured to:
- create a Finite State Machine (FSM) including a plurality of system states in consideration of dynamic factors that trigger a transition between system states for an original dataflow composed of a plurality of tasks, the dynamic factors indicating events that trigger the transition between the system states, wherein the dynamic factors are factors related to at least one of a Quality of Service (QoS) change and a power budget change, and wherein the plurality of system states are sequentially arranged in accordance with a QoS level or a power budget level;
- optimize the original dataflow through optimization of the created FSM and store the optimized original dataflow in a database;
- in response to detecting a change in system state during runtime, control an unmanned flying object to autonomically adapt by extracting a response to the change in system state on the basis of the optimized original dataflow stored in the database, and provide the response to the unmanned flying object,
- wherein the processor is further configured to create the FSM such that a state transition is possible only between adjacent system states among the plurality of system states that were sequentially arranged,
- wherein the processor is further configured to:
- map each of the plurality of tasks corresponding to the created FSM to at least one core of multiple cores to decide candidate mapping information;
- and
- determine whether predetermined conditions are satisfied by applying a priority-based list-scheduling to the plurality of tasks mapped to the multiple cores, and decide the candidate mapping information as a final mapping information when the predetermined conditions are satisfied as a result of the determination,
- wherein in the mapping, each task of the plurality of tasks is designated with at least one slot among slots of the multiple cores as many as a maximum parallelism degree of the each task of the plurality of tasks, and
- wherein the predetermined conditions include a condition that sets latency to be a predetermined time or less in parallelizing the plurality of tasks scheduled through the multiple cores.
11 more elements
Dependencies are inferred from the claim text.