Stratified sampling for even workload partitioning

This work presents a novel algorithm, Workload Partitioning and Scheduling (WPS), for evenly partitioning the computational workload of large implicitly-defined work-list based applications on distributed/shared-memory systems. WPS uses stratified sampling to estimate the number of work items that w...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	PACT '14 : proceedings of the 23rd International Conference on Parallel Architectures and Compilation Techniques : August 24-27, 2014, Edmonton, AB, Canada s. 503 - 504
Hlavní autoři:	Paudel, Jeeva, Amaral, Jose Nelson
Médium:	Konferenční příspěvek
Jazyk:	angličtina
Vydáno:	ACM 01.08.2014
Témata:	APGAS Clustering algorithms Load balancing Load management Partitioning algorithms PGAS Processor scheduling Programming Scheduling Stratified sampling X10
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	This work presents a novel algorithm, Workload Partitioning and Scheduling (WPS), for evenly partitioning the computational workload of large implicitly-defined work-list based applications on distributed/shared-memory systems. WPS uses stratified sampling to estimate the number of work items that will be processed in each step of an application. WPS uses such estimation to evenly partition and distribute the computational workload. An empirical evaluation on large applications - Iterative-Deepening A* (IDA*) applied to (4×4)-Sliding-Tile Puzzles, Delaunay Mesh Generation, and Delaunay Mesh Refinement - shows that WPS is applicable to a range of problems, and yields 28% to 49% speedups over existing work-stealing schedulers alone.
DOI:	10.1145/2628071.2671422