Title | ||
---|---|---|
MOSAIC: Heterogeneity-, Communication-, and Constraint-Aware Model Slicing and Execution for Accurate and Efficient Inference |
Abstract | ||
---|---|---|
Heterogeneous embedded systems have surfaced as a promising solution for accurate and efficient deep-learning inference on mobile devices. Despite extensive prior works, it still remains unexplored to investigate the system-software support that efficiently executes inference workloads by judiciously considering their performance and energy heterogeneity, communication overheads, and constraints. To bridge this gap, we propose MOSAIC, heterogeneity-, communication-, and constraint-aware model slicing and execution for accurate and efficient inference on heterogeneous embedded systems. MOSAIC generates the efficient model slicing and execution plan for the target inference workload through dynamic programming. MOSAIC significantly reduces inference latency and energy, exhibits high estimation accuracy, and incurs small overheads. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/PACT.2019.00021 | 2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT) |
Keywords | Field | DocType |
Model Slicing and Execution,Inference,Heterogeneous Embedded Systems | Dynamic programming,Workload,Inference,Latency (engineering),Computer science,Slicing,Parallel computing,Mobile device,Memory management,Energy consumption | Conference |
ISSN | ISBN | Citations |
1089-795X | 978-1-7281-3614-1 | 1 |
PageRank | References | Authors |
0.34 | 14 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Myeonggyun Han | 1 | 9 | 3.85 |
Jihoon Hyun | 2 | 1 | 0.34 |
Seongbeom Park | 3 | 9 | 2.83 |
Jinsu Park | 4 | 36 | 7.43 |
Woongki Baek | 5 | 402 | 25.85 |