PolyGraph: Exposing the Value of Flexibility for Graph Processing Accelerators

Because of the importance of graph workloads and the limitations of CPUs/GPUs, many graph processing accelerators have been proposed. The basic approach of prior accelerators is to focus on a single graph algorithm variant (eg. bulk-synchronous + slicing). While helpful for specialization, this leav...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Proceedings - International Symposium on Computer Architecture s. 595 - 608
Hlavní autoři: Dadu, Vidushi, Liu, Sihao, Nowatzki, Tony
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.06.2021
Témata:
accelerators
caches
Computational modeling
dataflow
flexibility
graph processing
Heuristic algorithms
Memory management
Parallel processing
Processor scheduling
reconfigurable
slicing
Switches
synchronous
tasks
Taxonomy
work efficiency
ISSN:2575-713X
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Because of the importance of graph workloads and the limitations of CPUs/GPUs, many graph processing accelerators have been proposed. The basic approach of prior accelerators is to focus on a single graph algorithm variant (eg. bulk-synchronous + slicing). While helpful for specialization, this leaves performance potential from flexibility on the table and also complicates understanding the relationship between graph types, workloads, algorithms, and specialization.In this work, we explore the value of flexibility in graph processing accelerators. First, we identify a taxonomy of key algorithm variants. Then we develop a template architecture (PolyGraph) that is flexible across these variants while being able to modularly integrate specialization features for each.Overall we find that flexibility in graph acceleration is critical. If only one variant can be supported, asynchronous-updates/priority-vertex-scheduling/graph-slicing is the best design, achieving 1.93× speedup over the best-performing accelerator, GraphPulse. However, static flexibility per-workload can further improve performance by 2.71×. With dynamic flexibility per-phase, performance further improves by up to 50%.
ISSN:2575-713X
DOI:10.1109/ISCA52012.2021.00053