An efficient GPU-based parallel tabu search algorithm for hardware/software co-design

Hardware/software partitioning is an essential step in hardware/software co-design. For large size problems, it is difficult to consider both solution quality and time. This paper presents an efficient GPU-based parallel tabu search algorithm (GPTS) for HW/SW partitioning. A single GPU kernel of com...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Frontiers of Computer Science Jg. 14; H. 5; S. 145316
Hauptverfasser: HOU, Neng, HE, Fazhi, ZHOU, Yi, CHEN, Yilin
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Beijing Higher Education Press 01.10.2020
Springer Nature B.V
Schlagworte:
Co-design
Computer Science
Costs
Digital signal processors
Embedded systems
Energy consumption
Genetic algorithms
GPU-based parallel tabu search
graphics processing unit
Hardware
hardware/software co-design
hardware/software partitioning
kernel fusion strategy
Neighborhoods
Optimization
optimized transfer strategy
Partitioning
Research Article
Search algorithms
single kernel implementation
Software
Strategy
Tabu search
Traveling salesman problem
optimized transfer strategy
single kernel implementation
GPU-based parallel tabu search
graphics processing unit
hardware/software co-design
kernel fusion strategy
hardware/ software partitioning
ISSN:2095-2228, 2095-2236
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Hardware/software partitioning is an essential step in hardware/software co-design. For large size problems, it is difficult to consider both solution quality and time. This paper presents an efficient GPU-based parallel tabu search algorithm (GPTS) for HW/SW partitioning. A single GPU kernel of compacting neighborhood is proposed to reduce the amount of GPU global memory accesses theoretically. A kernel fusion strategy is further proposed to reduce the amount of GPU global memory accesses of GPTS. To further minimize the transfer overhead of GPTS between CPU and GPU, an optimized transfer strategy for GPU-based tabu evaluation is proposed, which considers that all the candidates do not satisfy the given constraint. Experiments show that GPTS outperforms state-of-the-art work of tabu search and is competitive with other methods for HW/SW partitioning. The proposed parallelization is significant when considering the ordinary GPU platform.
Bibliographie:optimized transfer strategy
single kernel implementation
Document accepted on :2019-08-12
GPU-based parallel tabu search
graphics processing unit
hardware/software co-design
hardware/software partitioning
kernel fusion strategy
Document received on :2018-05-14
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2095-2228
2095-2236
DOI:10.1007/s11704-019-8184-3