Efficient and Productive Parallel Data Processing in Hybrid FPGA-CPU Reconfigurable Clusters with Resilient Distributed Datasets
Grant Data
Project Title
Efficient and Productive Parallel Data Processing in Hybrid FPGA-CPU Reconfigurable Clusters with Resilient Distributed Datasets
Principal Investigator
Dr So, Hayden Kwok Hay
(Principal Investigator (PI))
Co-Investigator(s)
Professor Tsia Kevin Kin Man
(Co-Investigator)
Professor Wawrzynek John
(Co-Investigator)
Duration
48
Start Date
2017-01-01
Amount
675647
Conference Title
Efficient and Productive Parallel Data Processing in Hybrid FPGA-CPU Reconfigurable Clusters with Resilient Distributed Datasets
Presentation Title
Keywords
computer architecture, FPGA, heterogeneous computing, reconfigurable computing, resilient distributed dataset
Discipline
Computing Hardware,Software
Panel
Engineering (E)
HKU Project Code
17245716
Grant Type
General Research Fund (GRF)
Funding Year
2016
Status
Completed
Objectives
1) To evaluate and extend the use of RDD as the abstract model of computations in hybrid FPGA-CPU clusters; 2) To develop a high-level design and compilation framework for developing and deploying RDD processing applications on a hybrid FPGA-CPU cluster; 3) To investigate mechanisms and performance tradeoffs for just-in-time hardware compilation that utilizes an RDD processing FPGA overlay. 4) To develop run-time hybrid resource scheduling and mapping algorithms for effective hardware-software integration.